system

Abstract

A system is provided.SOLUTION: A system including means for acquiring posting data on an SNS, means for classifying the acquired posting data into a text, an image, and a moving image, means for analyzing the classified data by natural language processing, computer vision, and voice analysis, means for predicting a mental health risk from the analyzed data, means for generating a warning message and a corresponding message on the basis of a prediction result, and means for notifying a user of the generated message.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] In recent years, mental health problems have been increasing with the spread of social media. However, it is extremely difficult to detect mental health problems early from social media posts and provide appropriate support. Conventional methods have difficulty comprehensively analyzing the content of posts, and there are limited methods for integrating and analyzing data in different formats, such as text, images, and videos. As a result, signs of mental health problems are often overlooked, and problems often go unnoticed until they become serious. The present invention aims to solve these problems. [Means for solving the problem]

[0005] The present invention solves the above problems by the following means: Provides a means for acquiring posted data on SNS, and collects the data with the user's explicit permission; Provides a means for classifying the acquired posted data into text, images, and videos; Further, provides a means for analyzing this classified data using natural language processing, computer vision, and speech analysis; Provides a means for integrating the analyzed data and predicting the user's mental health risk using a generative AI model; Provides a system that provides a means for generating warning messages and response messages based on the prediction results and a means for notifying the user of these, thereby enabling early detection of mental health problems and prompt provision of appropriate support.

[0006] "SNS" is an abbreviation for social networking service, which refers to an online platform where users can share information and opinions over the Internet.

[0007] "Posted data" refers to digital content such as text, images, and videos that users make public or private on social media.

[0008] "Means of acquisition" refers to methods or systems for incorporating user-posted digital content into a server using SNS APIs or web crawling technology.

[0009] "Natural language processing" refers to the techniques and algorithms that enable computers to understand and process human language (especially text data).

[0010] "Computer vision" refers to the technologies and algorithms that allow computers to analyze, recognize, understand, and generate visual data such as images and videos.

[0011] "Speech analysis" refers to the technology and algorithms used to analyze voice data, extract the information contained within it, and understand it.

[0012] A "generative AI model" refers to an artificial intelligence model that uses machine learning and deep learning to generate new information and predictions from data.

[0013] "Mental health risk" refers to the possibility of a decline in a user's mental health, or indicators or conditions that indicate such a decline.

[0014] "Warning message" refers to a warning message sent to the user when the system detects a mental health risk.

[0015] "Response message" refers to a message that suggests specific actions that the user should take in response to detected mental health risks.

[0016] "Means of notifying the user" refers to the method or system for delivering the generated message to the user by means of email, push notification, in-app message, etc. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8]FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0018] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0019] First, the terms used in the following description will be explained.

[0020] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0021] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0022] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0023] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0024] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0025] [First embodiment]

[0026] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0027] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0028] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0029] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0030] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0031] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0032] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0033] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0034] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0035] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0036] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0037] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0038] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[0039] 1. Data Collection

[0040] The server uses the API of the SNS platform to collect user posted data. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[0041] 2. Data Classification

[0042] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[0043] 3. Data Analysis

[0044] The server analyzes the data using various analysis modules.

[0045] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[0046] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[0047] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[0048] 4. Mental health risk prediction

[0049] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[0050] 5. Generating warning and response messages

[0051] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0052] 6. User Notices

[0053] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[0054] Specific examples

[0055] Example 1: For user A

[0056] User A posts on social media, "I've been feeling really tired lately."

[0057] The server collects the posts and sends them to a text analysis module.

[0058] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0059] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0060] A generative AI model integrates this data and predicts elevated mental health risks.

[0061] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[0062] The terminal notifies this message to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[0063] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[0064] The processing flow will be explained below.

[0065] Step 1:

[0066] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[0067] Step 2:

[0068] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[0069] Step 3:

[0070] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[0071] Step 4:

[0072] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[0073] Step 5:

[0074] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral facial expressions or dark backgrounds.

[0075] Step 6:

[0076] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[0077] Step 7:

[0078] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[0079] Step 8:

[0080] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[0081] Step 9:

[0082] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[0083] Step 10:

[0084] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0085] Step 11:

[0086] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[0087] Step 12:

[0088] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[0089] Example 1

[0090] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0091] In recent years, with the spread of social networking sites, it has become increasingly difficult for users to recognize their own mental health risks. In particular, in today's world where posting on social networking sites has become a part of everyday life, technology that can analyze this posting data and detect users' mental health status early is crucial. However, existing systems lack the functionality to perform multifaceted analysis of posting data, making it difficult to accurately predict mental health risks and provide users with appropriate warnings and response messages.

[0092] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0093] In this invention, the server includes means for acquiring posted data on SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and video analysis, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages based on the prediction results, and means for notifying the generated messages to the user. This enables multifaceted analysis of the user's mental health status from SNS posted data, and enables highly accurate risk prediction and generation of warning and response messages.

[0094] "Means for obtaining data posted on SNS" refers to a function that uses the API of a social networking service (SNS) platform to collect data such as text, images, and videos posted with the user's permission.

[0095] "Means for classifying acquired posted data into text, images, and videos" is a function that detects the format of collected posted data and sorts it into the appropriate module as text data, image data, or video data.

[0096] "Means for analyzing classified data using natural language processing, computer vision, and video analysis" refers to functions for analyzing text data using natural language processing technology (NLP), analyzing image data using computer vision technology, and analyzing video data using video analysis technology.

[0097] "Means for integrating analyzed data and using a generative AI model to predict mental health risks" refers to a function that integrates analyzed text, image, and video data and uses a generative AI model to predict a user's mental health risks.

[0098] The "means for generating a warning message and a response message based on the prediction result" is a function for generating a warning message for the user and a message suggesting specific response methods based on the mental health risk prediction result.

[0099] The "means for notifying the user of the generated message" is a function for sending the generated warning message and corresponding message to the user via a dedicated app, email, or push notification.

[0100] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[0101] Hardware and software used

[0102] Hardware:

[0103] Server: High-performance server (e.g., AWS (registered trademark) EC2, Google (registered trademark) Cloud Compute Engine)

[0104] Device: Smartphone (e.g. iPhone (registered trademark), ANDROID (registered trademark) device)

[0105] software:

[0106] Natural Language Processing algorithms: BERT and GPT-based models

[0107] Computer vision technology: OpenCV, TENSORFLOW (registered trademark), Keras

[0108] Video analysis algorithms: OpenPose, YOLOv3

[0109] Generative AI models: GPT-3 (registered trademark) and other latest generative AI models

[0110] Data collection

[0111] The server collects user post data through the API of the social media platform. Specifically, the server periodically calls the API to retrieve user posts (text, images, videos). This also includes a step to obtain the user's explicit permission. For example, the server uses the Twitter API to call "GET statuses / user_timeline" to collect the latest posts from a specified user.

[0112] Data Classification

[0113] The server classifies the collected data into three modules: text, image, and video. Specifically, it checks the data format and sends the data to the appropriate analysis module. For example, text data is sent to the NLP (natural language processing) module, image data is sent to the computer vision module, and video data is sent to the video analysis module.

[0114] Data analysis

[0115] The server analyzes the data in each module. The specific operations are as follows:

[0116] Text data: The NLP module performs keyword extraction and sentiment analysis using BERT and GPT-based models. For example, it extracts the keyword "tired" and determines the sentiment as "fatigue."

[0117] Image data: The computer vision module uses OpenCV and TensorFlow to perform facial expression recognition, object recognition, and scene analysis, for example, analyzing whether a face in an image has a neutral expression.

[0118] Video data: The video analysis module uses speech recognition and video analysis algorithms (e.g., OpenPose, YOLOv3) to perform multifaceted sentiment analysis, such as analyzing changes in voice tone and facial expressions.

[0119] Mental health risk prediction

[0120] The server integrates the analysis results and uses a generative AI model to predict mental health risks. Specifically, it compares the user's past data with the current analysis results to detect abnormal emotional and behavioral patterns. For example, if the user's posted text frequently contains the word "tired," this will be added up as a risk.

[0121] Generates warnings and response messages

[0122] If the server predicts a high mental health risk, it immediately generates a warning message. Specifically, it generates messages such as "Consult a specialist" or "Take a rest" depending on the risk level. The generative AI model also suggests specific actions that the user can take. For example, it generates a message that reads, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist."

[0123] User Notifications

[0124] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and are selected based on the user's settings. For example, the push notification function of a smartphone app can be used to display a message saying "Consult an expert." However, email notification is also possible.

[0125] Specific examples

[0126] Example 1: For user A

[0127] User A posts on social media, "I've been feeling really tired lately."

[0128] The server collects these posts using the SNS API and sends them to the text analysis module.

[0129] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0130] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0131] A generative AI model integrates this data to predict elevated mental health risks.

[0132] The server displays a warning message saying, "Your mental health is at increased risk. Take a break or consult a professional," along with a suggestion for specific action: "Start practicing mindfulness to reduce stress."

[0133] The device sends this message as a push notification to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[0134] Prompt Sentence Examples

[0135] "Predict the mental health risk of users who recently posted on social media that they've been feeling very tired lately."

[0136] "Identify the mental health risk from the following post: 'Image: Expressionless face, Text: I've been feeling tired lately' and generate an appropriate warning message."

[0137] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks in users early from SNS posting data and respond promptly.

[0138] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0139] Step 1: Data collection

[0140] The server collects user post data using the API of the SNS platform. Specifically, the server periodically makes API calls to retrieve user posts (text, images, videos) using endpoints such as "GET statuses / user_timeline". The input is the post data obtained as a result of the SNS API call, and the output is the retrieved raw data.

[0141] Step 2: Data Classification

[0142] The server classifies the collected posted data into three types: text, image, and video. The server checks the data format and distributes text data to the NLP module, image data to the computer vision module, and video data to the video analysis module. The input is the collected raw data, and the output is data classified by format.

[0143] Step 3: Text data analysis

[0144] The server processes the text data in a text analysis module. Specifically, it uses BERT or GPT-based natural language processing (NLP) models to extract keywords and perform sentiment analysis. For example, from the text posted by a user saying "I've been feeling tired lately," it extracts the keyword "tired" and recognizes the emotion as "fatigue." The input is text data, and the output is the analyzed text data and extracted sentiment information.

[0145] Step 4: Image data analysis

[0146] The server processes image data using an image analysis module. Specifically, it uses OpenCV and TensorFlow to recognize facial expressions in images and determine their emotions. For example, it performs expressionless face detection and background scene analysis. The input is image data, and the output is analyzed image data and emotional information.

[0147] Step 5: Video data analysis

[0148] The server processes the video data using a video analysis module. Specifically, it uses OpenPose and YOLOv3 to analyze the audio and video in the video and detect emotional changes. For example, it analyzes changes in voice tone and facial expressions. The input is the video data, and the output is the analyzed video data and emotional information.

[0149] Step 6: Data integration and mental health risk prediction

[0150] The server integrates the analysis results obtained from text, images, and videos and uses a generative AI model to predict mental health risks. Specifically, it centralizes this data and compares it with the user's past data to detect abnormal patterns. The input is the analyzed text, image, and video data, and the output is the predicted mental health risk.

[0151] Step 7: Generate warning and response messages

[0152] The server generates a warning message based on the risk prediction result. For example, it generates a message such as "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist," or a message suggesting specific measures to take, such as "Take a break" or "Talk to a friend." The input is the risk prediction result, and the output is the generated warning message and response message.

[0153] Step 8: User Notification

[0154] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and notifications are sent in the appropriate way depending on the user's settings. For example, the push notification function of a smartphone app can be used to display the generated warning message. The input is the generated warning message and the corresponding message, and the output is the notification sent to the user.

[0155] (Application example 1)

[0156] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0157] In recent years, the spread of social media has led to many people sharing information on a daily basis. However, at the same time, there has been an increase in the number of mental health risks hidden in posts on social media, and there is a need to detect these risks early and take appropriate measures. In addition, there is a problem that it is difficult to respond in real time due to the lack of technology that can instantly notify users of risks using devices such as smartphones or smart glasses.

[0158] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0159] In this invention, the server includes means for acquiring posted data on the SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the generated messages to the user, and means for notifying the user of the warning messages via a smartphone, smart glasses, or similar device. This enables early detection of mental health risks from the content of posts on the SNS and notification to the user in real time.

[0160] "SNS" is an online platform that allows users to share their information and opinions over the Internet.

[0161] "Posted data" refers to digital content such as text, images, and videos posted by users on SNS.

[0162] "Text classification" is the process of analyzing acquired post data based on the content of the strings and sorting them into specific categories.

[0163] "Image classification" is a technology that analyzes images in acquired posting data and classifies them into specific categories based on their content.

[0164] "Video classification" is a technology that analyzes videos in the acquired posting data and classifies them based on their content and characteristics.

[0165] "Natural language processing" is a general term for algorithms and technologies that allow computers to understand and analyze human language.

[0166] "Computer vision" refers to the technology and algorithms used to analyze and understand image and video data.

[0167] "Voice analysis" is a technology that analyzes voice data in digital format and recognizes its content and characteristics.

[0168] "Mental health risks" refer to conditions or symptoms that may have a negative impact on mental health.

[0169] A "warning message" is alert information that notifies the user of a crisis or risk.

[0170] The "response message" is information that suggests specific actions or methods of response to the user who receives the warning message.

[0171] A "generative AI model" is a technology that uses an artificial intelligence model that learns from existing data and experience to predict the characteristics and risks of new data.

[0172] "Notification means" refers to the method or technique for notifying the user of the generated message.

[0173] A "smartphone" is a mobile phone device with advanced computing power and connectivity.

[0174] "Smart glasses" are wearable devices that can integrate and display real and digital information.

[0175] "Device" is a general term for electronic devices that have specific functions or roles.

[0176] The system for implementing the present invention includes a wide range of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. A specific method for implementing this system is described below.

[0177] 1. Data Collection

[0178] The server uses the API of the SNS platform to collect user posted data. This requires the user's explicit permission. The collected data is multimodal data such as text, images, and videos.

[0179] 2. Data Classification

[0180] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[0181] 3. Data Analysis

[0182] The server analyzes the data using various analysis modules.

[0183] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[0184] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[0185] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[0186] 4. Mental health risk prediction

[0187] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[0188] 5. Generating warning and response messages

[0189] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0190] 6. User Notices

[0191] The device notifies the user of the generated message via a dedicated app, email, or push notification. Additionally, notifications can be sent via smartphones, smart glasses, or similar devices. The user can then review the notifications to understand their mental health status and any necessary actions.

[0192] Specific examples

[0193] Example 1: For user A

[0194] User A posts on social media, "I've been feeling really tired lately." The server collects this post and sends it to a text analysis module. An NLP algorithm extracts the keyword "tired," and sentiment analysis determines the feeling as "fatigue." Computer vision analysis detects that the posted image contains an inconspicuous background and an expressionless face. A generative AI model integrates this data and predicts a high mental health risk. The server generates a warning message for User A saying, "Your mental health risk is increasing. Please take a rest or consult a specialist." The device notifies User A of this message. User A checks the notification and takes appropriate action, such as consulting a specialist.

[0195] Example prompts to be input to the generative AI model

[0196] "Analyze the following text and image posted by a user on social media and predict their mental health risk. Text: "I'm so sad", Image: <Image URL>"

[0197] This system makes it possible to detect mental health risks from social media posts early on and provide prompt responses.

[0198] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0199] Step 1:

[0200] Data collection

[0201] The server collects user posted data using the API of the SNS platform. First, with the user's explicit permission, it retrieves text, image, and video data from the SNS using a dedicated app and API key. The input is the user's SNS account information and API key, and the output is the collected SNS posted data.

[0202] Step 2:

[0203] Data Classification

[0204] The server classifies the collected data. The acquired text data is sent to the text analysis module, the image data is sent to the image analysis module, and the video data is sent to the video analysis module. The input is the collected SNS post data, and the output is classified text, image, and video data.

[0205] Step 3:

[0206] Text Data Analysis

[0207] The server analyzes the text data using natural language processing (NLP) algorithms. Specifically, it extracts keywords and performs sentiment analysis. The input is text data, and the output is analyzed emotional information and keywords. For example, the server can detect the emotion "sad" from the posted text "very sad."

[0208] Step 4:

[0209] Image data analysis

[0210] The server analyzes image data using computer vision technology. Specifically, it performs facial expression recognition, object recognition, and scene analysis. The input is image data, and the output is analyzed facial expression information and scene information. For example, neutral faces and dark backgrounds are detected.

[0211] Step 5:

[0212] Video Data Analysis

[0213] The server uses a video analysis algorithm to perform multifaceted emotion analysis using voice recognition and video analysis. The input is video data, and the output is analyzed emotional information and behavioral patterns. For example, if the user's voice tone is low and the video is dark, a report will be generated.

[0214] Step 6:

[0215] Data integration and risk prediction

[0216] The server integrates the analyzed text, image, and video data and uses a generative AI model to predict mental health risks. The input is multiple analyzed data (emotional information, facial expression information, and behavioral patterns), and the output is the predicted mental health risk. For example, abnormal emotional and behavioral patterns can be detected by comparing them with past data.

[0217] Step 7:

[0218] Generates warnings and response messages

[0219] The server generates a warning message and a corresponding response message based on the risk prediction result. The input is the mental health risk prediction result, and the output is the generated warning message and corresponding response message. For example, a message such as "Your mental health risk is increasing. Take a rest or consult a specialist" may be generated.

[0220] Step 8:

[0221] User Notifications

[0222] The device notifies the user of the generated message. The notification is via a smartphone, smart glasses, or similar device. The input is the generated message, and the output is a notification displayed on the user's device. The user reviews the notification and receives information about their mental health condition and how to respond.

[0223] This processing step makes it possible to detect mental health risks early from the content of social media posts and notify users in real time.

[0224] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0225] The system of the present invention is a system for analyzing data posted on social networking sites to predict and warn of mental health risks, and in particular has the function of recognizing users' emotions in detail by incorporating an emotion engine. Specific embodiments for implementing the present invention are described below.

[0226] 1. Data Collection

[0227] The server uses the API of the SNS platform to collect user posted data. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[0228] 2. Data Classification

[0229] The server categorizes the collected data into text, images, and videos, and sends the categorized data to the corresponding analysis module.

[0230] 3. Data Analysis

[0231] The server analyzes the data using various analysis modules.

[0232] The text data is analyzed using natural language processing (NLP) algorithms. Specifically, keyword extraction and sentiment analysis are performed to detect, for example, the keyword "fatigue" and its associated sentiment (fatigue).

[0233] The image data is then subjected to computer vision techniques for facial expression recognition, object recognition, and scene analysis, for example to detect neutral faces and dark backgrounds.

[0234] Video data is analyzed using a video analysis algorithm, which uses voice recognition and video analysis to perform multifaceted emotion analysis, such as detecting low voice tones and little movement.

[0235] 4. Emotion Recognition by Emotion Engine

[0236] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[0237] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions.

[0238] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions, to identify the emotional state in the image.

[0239] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc.

[0240] 5. Mental health risk prediction

[0241] The server integrates the emotional data recognized by the emotion engine and predicts mental health risks using a generative AI model, which compares the data with the user's past data to detect abnormal emotional and behavioral patterns.

[0242] 6. Generating warning and response messages

[0243] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0244] 7. User Notices

[0245] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[0246] Specific examples

[0247] Example 1: For user A

[0248] User A posts on social media, "I've been feeling really tired lately."

[0249] The server collects the posts and sends them to a text analysis module.

[0250] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0251] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0252] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[0253] A generative AI model integrates this data and predicts elevated mental health risks.

[0254] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[0255] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[0256] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[0257] The processing flow will be explained below.

[0258] Step 1:

[0259] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[0260] Step 2:

[0261] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[0262] Step 3:

[0263] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[0264] Step 4:

[0265] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[0266] Step 5:

[0267] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral facial expressions or dark backgrounds.

[0268] Step 6:

[0269] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[0270] Step 7:

[0271] Emotion recognition by emotion engine. The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[0272] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions, for example, recognizing negative emotions from the phrase "tired."

[0273] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions.

[0274] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc. For example, a low voice tone and little movement can indicate fatigue.

[0275] Step 8:

[0276] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[0277] Step 9:

[0278] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[0279] Step 10:

[0280] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[0281] Step 11:

[0282] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0283] Step 12:

[0284] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[0285] Step 13:

[0286] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[0287] Example 2

[0288] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0289] Conventional systems have had difficulty accurately predicting users' mental health risks from data posted on social media and generating appropriate warning and response messages. Furthermore, it is not easy to comprehensively analyze a variety of data (text, images, and videos) with explicit permission from the user. This makes it difficult to address users' mental health issues early, and there is a need for improved accuracy in risk prediction.

[0290] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0291] In this invention, the server includes means for acquiring data posted on SNS, means for classifying the posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for performing detailed emotion recognition on the analyzed text, image, and video data using an emotion engine, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages, and means for notifying the generated messages to the user. This makes it possible to accurately predict a user's mental health risks from data posted on SNS and provide early and appropriate responses.

[0292] "SNS posting data" refers to content such as text, images, and videos posted by users on social networking services (SNS).

[0293] "Text" refers to data containing text information posted by users on SNS.

[0294] "Images" are data containing visual content posted by users on social networking sites, including photographs and illustrations.

[0295] "Video" refers to data containing content accompanied by motion and sound that users post on social networking sites.

[0296] "Natural language processing" is a technology that uses computers to analyze and understand human language, and includes text data analysis, keyword extraction, sentiment analysis, and more.

[0297] "Computer vision" is a technology that allows computers to analyze and understand visual information such as images and videos.

[0298] "Voice analysis" is a technology that analyzes voice data and performs functions such as converting it into text, analyzing emotions, and recognizing speech.

[0299] An "emotion engine" is a technology that comprehensively analyzes data such as text, images, and videos to recognize the user's emotional state in detail.

[0300] A "generative AI model" is an artificial intelligence model that uses machine learning technology to recognize specific patterns from input data and make predictions or generate results.

[0301] "Mental health risk" refers to the possibility that a user may have problems with their mental health, and is predicted based on abnormal emotions and behavioral patterns from analytical data.

[0302] A "warning message" is a message that notifies the user that their mental health risk is increasing.

[0303] The "response message" is a message that suggests specific ways to respond to a user whose mental health risk is increasing.

[0304] "Means of notification" refers to technology for transmitting the generated warning message and response message to the user, and includes dedicated apps, email, push notifications, etc.

[0305] The system of the present invention analyzes data posted on social media platforms to predict and warn of mental health risks. Specifically, the server uses the social media platform's API to collect user posted data (text, images, and videos), and classifies and analyzes this data to predict mental health risks. Furthermore, an emotion engine is incorporated to recognize users' emotions in detail, and appropriate warning and response messages are generated and sent.

[0306] Data collection

[0307] The server collects user posting data using the APIs of social networking platforms such as Twitter and Facebook. Multimodal data such as text, images, and videos are collected with the user's explicit permission. For example, if a user posts "I've been feeling tired lately," the text data and attached image and video data are also collected.

[0308] Data Classification

[0309] The server categorizes the collected posting data into three types: text, images, and videos. For example, a photo of a non-smiling face posted along with a text message saying "I'm not feeling well" or a video in a dark room would be classified separately. This ensures that data is properly sorted according to each analysis module.

[0310] Data analysis

[0311] The server analyzes the data using the following analysis modules:

[0312] The text data is then subjected to keyword extraction and sentiment analysis using NLP (Natural Language Processing) algorithms, for example, to detect the keyword "fatigue" and the associated "feeling of fatigue."

[0313] The image data is then subjected to facial expression recognition and scene analysis using computer vision techniques, such as detecting neutral facial expressions and dark backgrounds as features.

[0314] Video data is analyzed using voice recognition and video analysis to detect voice tone and lack of movement, for example, low voice tone and lack of movement.

[0315] Emotion recognition by emotion engine

[0316] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[0317] From the text data, phrase analysis is performed to recognize positive, negative, and neutral emotions. For example, a post saying "I've been feeling tired lately" can be interpreted as a negative emotion.

[0318] Emotions are recognized from image data by analyzing facial expressions and scenes, for example, detecting expressionless faces and dark scenes.

[0319] Video data is analyzed for audio and video to detect low tones and little movement.

[0320] Mental health risk prediction

[0321] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risk. This model compares a user's past emotional data with their current data to detect abnormal emotional and behavioral patterns. For example, if a user's recent posts show an abnormally negative trend compared to past data, it predicts a high mental health risk.

[0322] Generates warnings and response messages

[0323] If the server predicts a high mental health risk, it generates a warning message and a corresponding message, such as "Your mental health is at increased risk. Take a rest or consult a professional."

[0324] User Notifications

[0325] The device will notify the user of the warning message and response message sent from the server via a dedicated app, email, or push notification. Users can check the notification, understand their mental health risks, and take necessary actions.

[0326] Specific examples

[0327] Example 1: For user A

[0328] User A posts on social media, "I've been feeling really tired lately."

[0329] The server collects the posts and sends them to a text analysis module.

[0330] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0331] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0332] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[0333] A generative AI model integrates this data and predicts elevated mental health risks.

[0334] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[0335] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[0336] Example prompts for generative AI models

[0337] "Please tell me the system procedure for analyzing a user's emotional state from social media posts and predicting mental health risks."

[0338] "Please explain each processing step of the system that analyzes emotions based on user posts on social media and predicts mental health risks."

[0339] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks in users from posts on social media at an early stage and provide appropriate responses promptly.

[0340] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0341] Step 1: Data collection

[0342] The server uses the API of the SNS platform to collect user posted data. API authentication information (API key and secret) and user ID are required as input. The acquired data includes text, images, and videos. Specifically, the server sends an API request to the SNS platform and stores the user posted data in a temporary database.

[0343] Step 2: Data Classification

[0344] The server classifies the collected posted data into text, images, and videos. The input requires the multimodal data collected in step 1. The output is data classified into text data, image data, and video data. Specifically, the server checks the data format and distributes the data to the appropriate analysis module.

[0345] Step 3: Data analysis

[0346] The server analyzes data using various analysis modules. Text data, image data, and video data are required as input. Analysis results (keywords, emotional state, facial expression, voice tone, etc.) are obtained as output. The specific operation is as follows:

[0347] NLP algorithms are used to extract keywords and analyze sentiment from text data. For example, the keyword "tired" is extracted and the sentiment "fatigue" is determined.

[0348] Facial expression recognition is performed on image data using computer vision techniques. For example, neutral faces are detected.

[0349] Speech recognition and video analysis are performed on video data to detect voice tones and lack of movement. Example: Low voice tones and lack of movement are detected.

[0350] Step 4: Emotion Recognition with the Emotion Engine

[0351] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The analyzed data from step 3 is required as input. The output is a detailed emotional state (positive, negative, neutral). Specifically, the emotion engine performs phrase analysis, facial expression analysis, and voice analysis to recognize emotions from each media.

[0352] Step 5: Mental health risk prediction

[0353] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risks. Detailed emotional data and past emotional data are required as input. The output is a predicted mental health risk. Specifically, the system inputs emotional data into the generative AI model and performs risk assessment. In particular, it compares past data with current data to detect abnormal emotional and behavioral patterns.

[0354] Step 6: Generate warning and response messages

[0355] The server generates a warning message and a response message if a high mental health risk is predicted. The required input is the risk assessment result. The output is a warning message and a response message. Specifically, based on the risk assessment, the server generates a warning message such as "Take a rest or consult a specialist" and a message suggesting specific response methods such as "Talk to a friend" or "Consult a doctor."

[0356] Step 7: User Notification

[0357] The device notifies the user of the warning message and response message sent from the server. The generated message is required as input. The output is the completion of the notification to the user. The specific operation is to display the message via a dedicated app, email, or push notification. The user can then check the notification and take appropriate action.

[0358] (Application example 2)

[0359] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0360] In modern society, with the spread of self-driving vehicles, mental health issues among drivers are on the rise, and there is a need for a system that can monitor their mental state in real time and encourage appropriate responses. However, conventional technologies predict mental health risks solely from data posted on social media, making it difficult to reflect real-time conditions while driving. In addition, warning messages are not generated and notified quickly, making it difficult for drivers to recognize problems in a timely manner and take appropriate measures. These issues need to be resolved.

[0361] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring posted data on an SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and speech analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the user of the generated messages, means for analyzing data collected from sensors in the autonomous vehicle and evaluating the driver's mental state, and means for displaying a warning message using the vehicle's dashboard display system based on the evaluation results. This enables real-time monitoring of the driver's mental state while driving, and prompt warnings and response suggestions.

[0362] "SNS" is an abbreviation for social networking service, a platform for users to communicate and share information online.

[0363] "Posted data" refers to all information, including text, images, and videos, posted by users on SNS.

[0364] A "server" is a computer system that processes and manages information on a network.

[0365] "Natural language processing" refers to techniques and approaches that enable computers to understand and generate human language.

[0366] "Computer vision" is a technology that allows computers to analyze digital images and videos and understand visual information.

[0367] "Voice analysis" is a technology that analyzes voice data and extracts its content and characteristics.

[0368] "Mental health risk" refers to the possibility that an individual's mental health will deteriorate.

[0369] A "warning message" is a message that the system uses to notify the user of risks or situations that require caution.

[0370] A "response message" is a notification message that suggests specific actions or countermeasures for responding to a warning.

[0371] An "autonomous vehicle" is a vehicle that can be driven and navigated automatically without the need for human operation.

[0372] A "sensor" is a device that detects changes in the physical environment and acquires that information as electronic data.

[0373] A "dashboard" is the part of a vehicle that contains a display and interface through which the driver can view information.

[0374] This invention is a system that analyzes data posted on social media to predict and warn drivers about mental health risks, and is particularly applicable to self-driving vehicles. The system is configured as follows:

[0375] Hardware

[0376] Camera: Installed inside the vehicle to capture the driver's facial expressions and movements.

[0377] Microphone: Installed inside the vehicle to collect the driver's voice and audio.

[0378] Smart glasses: Worn by the driver, they track eye movements and gaze in real time.

[0379] Dashboard display: A display for showing warning and response messages.

[0380] software

[0381] SNS API: Retrieves post data from social media platforms.

[0382] Natural language processing (NLP) engine: Analyzes text data, extracts keywords, and performs sentiment analysis.

[0383] Computer vision tools: Analyze image data and perform facial expression and object recognition.

[0384] Video analysis algorithm: Analyzes video data and detects emotions from audio and video.

[0385] Emotion engine: Integrates various analytical data to recognize the driver's emotional state in detail.

[0386] Generative AI model: Predicting mental health risks based on integrated emotional data.

[0387] Notification system: Notifies the driver of warning and response messages via the dashboard display.

[0388] Program processing

[0389] The server collects driver posting data using SNS APIs. The collected data is categorized into text, images, and videos, and each data is sent to the corresponding analysis module. The NLP engine analyzes the text data to extract keywords and perform sentiment analysis. The computer vision tool analyzes the image data to recognize facial expressions and scenes. The video analysis algorithm analyzes the video data and extracts emotions from audio and video.

[0390] The analyzed data is integrated into an emotion engine to recognize the driver's emotional state in detail. The generative AI model uses this emotional data to predict mental health risks and immediately generates a warning message if the risk is high. For example, it might say, "Your mental health risk is increasing. Please take a break or consult a specialist." It also generates a response message suggesting specific ways to respond.

[0391] These messages are displayed on the dashboard display, allowing the driver to check their own mental health status in real time and take necessary action. This system makes it possible to detect mental health risks while driving early and provide appropriate responses quickly.

[0392] Specific examples

[0393] For example, if Driver B posts on social media while driving, "I'm tired right now, but I'm driving hard," the server collects this post, extracts the keyword "tired" using an NLP engine, and determines the state as "fatigue" using sentiment analysis. Furthermore, the vehicle's camera detects expressionless faces, and the microphone recognizes monotone voices. This data is integrated, and the sentiment engine precisely recognizes the emotions of "fatigue" and "expressionless," and the generative AI model predicts a high mental health risk. The system generates a warning message saying, "Your mental health risk is increasing. You should take a break," and displays this on the dashboard display. Driver B sees this message and takes action to rest in a safe place.

[0394] Prompt Sentence Examples

[0395] Input prompt: "Based on recent posts, predict risks and generate a warning message."

[0396] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0397] Step 1:

[0398] The server collects driver posting data using SNS API. The collected data is in the form of text, images, and videos, and is executed with the user's explicit permission. The SNS posting data is used as input, and it is output in the form of classified data.

[0399] Step 2:

[0400] The server classifies the collected data into text, images, and videos. The input here is the various posted data collected in step 1, and the classified data format (text, images, videos) is output. This classification is done automatically.

[0401] Step 3:

[0402] The server analyzes the classified data. Specifically, text data is analyzed using a natural language processing (NLP) engine to extract keywords and perform sentiment analysis. Image data is analyzed using computer vision tools to recognize facial expressions and scenes. Video data is analyzed using video analysis algorithms to extract emotions from audio and video. The input is the data classified in step 2, and the analyzed emotional data is the output.

[0403] Step 4:

[0404] The server integrates the analyzed data and performs detailed emotion recognition using an emotion engine. Here, the analysis results of each piece of text, image, and video data are input, and the integrated emotion data is output. The emotion engine recognizes the emotional state of each piece of data in detail and provides integrated emotion information.

[0405] Step 5:

[0406] The server uses a generative AI model based on the integrated emotion data to predict mental health risks. The input is the integrated emotion data, and the output is the predicted mental health risk. This prediction is made by comparing it with past data to detect abnormal emotion and behavior patterns.

[0407] Step 6:

[0408] If the server predicts a high mental health risk, it immediately generates a warning message and a corresponding message. The input here is the predicted mental health risk, and the message output is something like "Your mental health risk is increasing. Take a rest or consult a specialist."

[0409] Step 7:

[0410] The terminal notifies the driver of the generated message. Specifically, it displays warning messages and response messages in real time on the dashboard display so that the driver can check them. The input is the message generated in step 6, and the output is the displayed message. The driver can check this message and take safety measures.

[0411] For example, if Driver B posts on social media, "I'm tired right now, but I'm driving hard," the camera will detect an expressionless face and the microphone will detect a monotone voice. These data will be integrated to recognize the emotions of "fatigue" and "expressionless," and the generative AI model will predict an increase in mental health risk and display an appropriate warning message.

[0412] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0413] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0414] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0415] [Second embodiment]

[0416] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0417] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0418] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0419] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0420] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0421] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0422] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0423] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0424] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0425] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0426] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0427] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0428] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[0429] 1. Data Collection

[0430] The server collects user posted data using the API of the SNS platform. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[0431] 2. Data Classification

[0432] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[0433] 3. Data Analysis

[0434] The server analyzes the data using various analysis modules.

[0435] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[0436] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[0437] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[0438] 4. Mental health risk prediction

[0439] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[0440] 5. Generating warning and response messages

[0441] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0442] 6. User Notices

[0443] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[0444] Specific examples

[0445] Example 1: For user A

[0446] User A posts on social media, "I've been feeling really tired lately."

[0447] The server collects the posts and sends them to a text analysis module.

[0448] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0449] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0450] A generative AI model integrates this data and predicts elevated mental health risks.

[0451] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[0452] The terminal notifies this message to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[0453] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[0454] The processing flow will be explained below.

[0455] Step 1:

[0456] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[0457] Step 2:

[0458] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[0459] Step 3:

[0460] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[0461] Step 4:

[0462] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[0463] Step 5:

[0464] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral expressions or dark backgrounds.

[0465] Step 6:

[0466] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[0467] Step 7:

[0468] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[0469] Step 8:

[0470] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[0471] Step 9:

[0472] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[0473] Step 10:

[0474] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0475] Step 11:

[0476] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[0477] Step 12:

[0478] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[0479] Example 1

[0480] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0481] In recent years, with the spread of social networking sites, it has become increasingly difficult for users to recognize their own mental health risks. In particular, in today's world where posting on social networking sites has become a part of everyday life, technology that can analyze this posting data and detect users' mental health status early is crucial. However, existing systems lack the functionality to perform multifaceted analysis of posting data, making it difficult to accurately predict mental health risks and provide users with appropriate warnings and response messages.

[0482] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0483] In this invention, the server includes means for acquiring posted data on SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and video analysis, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages based on the prediction results, and means for notifying the generated messages to the user. This enables multifaceted analysis of the user's mental health status from SNS posted data, and enables highly accurate risk prediction and generation of warning and response messages.

[0484] "Means for obtaining data posted on SNS" refers to a function that uses the API of a social networking service (SNS) platform to collect data such as text, images, and videos posted with the user's permission.

[0485] "Means for classifying acquired posted data into text, images, and videos" is a function that detects the format of collected posted data and sorts it into the appropriate module as text data, image data, or video data.

[0486] "Means for analyzing classified data using natural language processing, computer vision, and video analysis" refers to functions for analyzing text data using natural language processing technology (NLP), analyzing image data using computer vision technology, and analyzing video data using video analysis technology.

[0487] "Means for integrating analyzed data and using a generative AI model to predict mental health risks" refers to a function that integrates analyzed text, image, and video data and uses a generative AI model to predict a user's mental health risks.

[0488] The "means for generating a warning message and a response message based on the prediction result" is a function for generating a warning message for the user and a message suggesting specific response methods based on the mental health risk prediction result.

[0489] The "means for notifying the user of the generated message" is a function for sending the generated warning message and corresponding message to the user via a dedicated app, email, or push notification.

[0490] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[0491] Hardware and software used

[0492] Hardware:

[0493] Server: High-performance server (e.g. AWS EC2, Google Cloud Compute Engine)

[0494] Device: Smartphone (e.g. iPhone, Android device)

[0495] software:

[0496] Natural Language Processing algorithms: BERT and GPT-based models

[0497] Computer Vision Technologies: OpenCV, TensorFlow, Keras

[0498] Video analysis algorithms: OpenPose, YOLOv3

[0499] Generative AI models: GPT-3 and the latest generative AI models

[0500] Data collection

[0501] The server collects user post data through the API of the social media platform. Specifically, the server periodically calls the API to retrieve user posts (text, images, videos). This also includes a step to obtain the user's explicit permission. For example, the server uses the Twitter API to call "GET statuses / user_timeline" to collect the latest posts from a specified user.

[0502] Data Classification

[0503] The server classifies the collected data into three modules: text, image, and video. Specifically, it checks the data format and sends the data to the appropriate analysis module. For example, text data is sent to the NLP (natural language processing) module, image data is sent to the computer vision module, and video data is sent to the video analysis module.

[0504] Data analysis

[0505] The server analyzes the data in each module. The specific operations are as follows:

[0506] Text data: The NLP module performs keyword extraction and sentiment analysis using BERT and GPT-based models. For example, it extracts the keyword "tired" and determines the sentiment as "fatigue."

[0507] Image data: The computer vision module uses OpenCV and TensorFlow to perform facial expression recognition, object recognition, and scene analysis, for example, analyzing whether a face in an image has a neutral expression.

[0508] Video data: The video analysis module uses speech recognition and video analysis algorithms (e.g., OpenPose, YOLOv3) to perform multifaceted sentiment analysis, such as analyzing changes in voice tone and facial expressions.

[0509] Mental health risk prediction

[0510] The server integrates the analysis results and uses a generative AI model to predict mental health risks. Specifically, it compares the user's past data with the current analysis results to detect abnormal emotional and behavioral patterns. For example, if the user's posted text frequently contains the word "tired," this will be added up as a risk.

[0511] Generates warnings and response messages

[0512] If the server predicts a high mental health risk, it immediately generates a warning message. Specifically, it generates messages such as "Consult a specialist" or "Take a rest" depending on the risk level. The generative AI model also suggests specific actions that the user can take. For example, it generates a message that reads, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist."

[0513] User Notifications

[0514] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and are selected based on the user's settings. For example, the push notification function of a smartphone app can be used to display a message saying "Consult an expert." However, email notification is also possible.

[0515] Specific examples

[0516] Example 1: For user A

[0517] User A posts on social media, "I've been feeling really tired lately."

[0518] The server collects these posts using the SNS API and sends them to the text analysis module.

[0519] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0520] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0521] A generative AI model integrates this data to predict elevated mental health risks.

[0522] The server displays a warning message saying, "Your mental health is at increased risk. Take a rest or consult a professional," along with a suggestion for specific action: "Start practicing mindfulness to reduce stress."

[0523] The device sends this message as a push notification to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[0524] Prompt Sentence Examples

[0525] "Predict the mental health risk of a user who recently posted on social media, 'I've been feeling really tired lately.'"

[0526] "Identify the mental health risk from the following post: 'Image: Expressionless face, Text: I've been feeling tired lately' and generate an appropriate warning message."

[0527] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks in users early from SNS posting data and respond promptly.

[0528] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0529] Step 1: Data collection

[0530] The server collects user post data using the API of the SNS platform. Specifically, the server periodically makes API calls to retrieve user posts (text, images, videos) using endpoints such as "GET statuses / user_timeline". The input is the post data obtained as a result of the SNS API call, and the output is the retrieved raw data.

[0531] Step 2: Data Classification

[0532] The server classifies the collected posted data into three types: text, image, and video. The server checks the data format and distributes text data to the NLP module, image data to the computer vision module, and video data to the video analysis module. The input is the collected raw data, and the output is data classified by format.

[0533] Step 3: Text data analysis

[0534] The server processes the text data in a text analysis module. Specifically, it uses BERT or GPT-based natural language processing (NLP) models to extract keywords and perform sentiment analysis. For example, from the text posted by a user saying "I've been feeling tired lately," it extracts the keyword "tired" and recognizes the emotion as "fatigue." The input is text data, and the output is the analyzed text data and extracted sentiment information.

[0535] Step 4: Image data analysis

[0536] The server processes image data using an image analysis module. Specifically, it uses OpenCV and TensorFlow to recognize facial expressions in images and determine their emotions. For example, it performs expressionless face detection and background scene analysis. The input is image data, and the output is analyzed image data and emotional information.

[0537] Step 5: Video data analysis

[0538] The server processes the video data using a video analysis module. Specifically, it uses OpenPose and YOLOv3 to analyze the audio and video in the video and detect emotional changes. For example, it analyzes changes in voice tone and facial expressions. The input is the video data, and the output is the analyzed video data and emotional information.

[0539] Step 6: Data integration and mental health risk prediction

[0540] The server integrates the analysis results obtained from text, images, and videos and uses a generative AI model to predict mental health risks. Specifically, it centralizes this data and compares it with the user's past data to detect abnormal patterns. The input is the analyzed text, image, and video data, and the output is the predicted mental health risk.

[0541] Step 7: Generate warning and response messages

[0542] The server generates a warning message based on the risk prediction result. For example, it generates a message such as "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist," or a message suggesting specific measures to take, such as "Take a break" or "Talk to a friend." The input is the risk prediction result, and the output is the generated warning message and response message.

[0543] Step 8: User Notification

[0544] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and notifications are sent in the appropriate way depending on the user's settings. For example, the push notification function of a smartphone app can be used to display the generated warning message. The input is the generated warning message and the corresponding message, and the output is the notification sent to the user.

[0545] (Application example 1)

[0546] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0547] In recent years, the spread of social media has led to many people sharing information on a daily basis. However, at the same time, there has been an increase in the number of mental health risks hidden in posts on social media, and there is a need to detect these risks early and take appropriate measures. In addition, there is a problem that it is difficult to respond in real time due to the lack of technology that can instantly notify users of risks using devices such as smartphones or smart glasses.

[0548] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0549] In this invention, the server includes means for acquiring posted data on the SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the generated messages to the user, and means for notifying the user of the warning messages via a smartphone, smart glasses, or similar device. This enables early detection of mental health risks from the content of posts on the SNS and notification to the user in real time.

[0550] "SNS" is an online platform that allows users to share their information and opinions over the Internet.

[0551] "Posted data" refers to digital content such as text, images, and videos posted by users on SNS.

[0552] "Text classification" is the process of analyzing acquired post data based on the content of the strings and sorting them into specific categories.

[0553] "Image classification" is a technology that analyzes images in acquired posting data and classifies them into specific categories based on their content.

[0554] "Video classification" is a technology that analyzes videos in the acquired posting data and classifies them based on their content and characteristics.

[0555] "Natural language processing" is a general term for algorithms and technologies that allow computers to understand and analyze human language.

[0556] "Computer vision" refers to the technology and algorithms used to analyze and understand image and video data.

[0557] "Voice analysis" is a technology that analyzes voice data in digital format and recognizes its content and characteristics.

[0558] "Mental health risks" refer to conditions or symptoms that may have a negative impact on mental health.

[0559] A "warning message" is alert information that notifies the user of a crisis or risk.

[0560] The "response message" is information that suggests specific actions or methods of response to the user who has received the warning message.

[0561] A "generative AI model" is a technology that uses an artificial intelligence model that learns from existing data and experience to predict the characteristics and risks of new data.

[0562] "Notification means" refers to the method or technique for notifying the user of the generated message.

[0563] A "smartphone" is a mobile phone device with advanced computing power and connectivity.

[0564] "Smart glasses" are wearable devices that can integrate and display real and digital information.

[0565] "Device" is a general term for electronic devices that have specific functions or roles.

[0566] The system for implementing the present invention includes a wide range of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. A specific method for implementing this system is described below.

[0567] 1. Data Collection

[0568] The server uses the API of the SNS platform to collect user posted data. This requires the user's explicit permission. The collected data is multimodal data such as text, images, and videos.

[0569] 2. Data Classification

[0570] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[0571] 3. Data Analysis

[0572] The server analyzes the data using various analysis modules.

[0573] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[0574] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[0575] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[0576] 4. Mental health risk prediction

[0577] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[0578] 5. Generating warning and response messages

[0579] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0580] 6. User Notices

[0581] The device notifies the user of the generated message via a dedicated app, email, or push notification. Additionally, notifications can be sent via smartphones, smart glasses, or similar devices. The user can then review the notifications to understand their mental health status and any necessary actions.

[0582] Specific examples

[0583] Example 1: For user A

[0584] User A posts on social media, "I've been feeling really tired lately." The server collects this post and sends it to a text analysis module. An NLP algorithm extracts the keyword "tired," and sentiment analysis determines the feeling as "fatigue." Computer vision analysis detects that the posted image contains an inconspicuous background and an expressionless face. A generative AI model integrates this data and predicts a high mental health risk. The server generates a warning message for User A saying, "Your mental health risk is increasing. Please take a rest or consult a specialist." The device notifies User A of this message. User A checks the notification and takes appropriate action, such as consulting a specialist.

[0585] Example prompts to be input to the generative AI model

[0586] "Analyze the following text and image posted by a user on social media and predict their mental health risk. Text: "I'm so sad", Image: <Image URL>"

[0587] This system makes it possible to detect mental health risks from social media posts early on and provide prompt responses.

[0588] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0589] Step 1:

[0590] Data collection

[0591] The server collects user posted data using the API of the SNS platform. First, with the user's explicit permission, it retrieves text, image, and video data from the SNS using a dedicated app and API key. The input is the user's SNS account information and API key, and the output is the collected SNS posted data.

[0592] Step 2:

[0593] Data Classification

[0594] The server classifies the collected data. The acquired text data is sent to the text analysis module, the image data is sent to the image analysis module, and the video data is sent to the video analysis module. The input is the collected SNS post data, and the output is classified text, image, and video data.

[0595] Step 3:

[0596] Text Data Analysis

[0597] The server analyzes the text data using natural language processing (NLP) algorithms. Specifically, it extracts keywords and performs sentiment analysis. The input is text data, and the output is analyzed emotional information and keywords. For example, the server can detect the emotion "sad" from the posted text "very sad."

[0598] Step 4:

[0599] Image data analysis

[0600] The server analyzes image data using computer vision technology. Specifically, it performs facial expression recognition, object recognition, and scene analysis. The input is image data, and the output is analyzed facial expression information and scene information. For example, neutral faces and dark backgrounds are detected.

[0601] Step 5:

[0602] Video Data Analysis

[0603] The server uses a video analysis algorithm to perform multifaceted emotion analysis using voice recognition and video analysis. The input is video data, and the output is analyzed emotional information and behavioral patterns. For example, if the user's voice tone is low and the video is dark, a report will be generated.

[0604] Step 6:

[0605] Data integration and risk prediction

[0606] The server integrates the analyzed text, image, and video data and uses a generative AI model to predict mental health risks. The input is multiple analyzed data (emotional information, facial expression information, and behavioral patterns), and the output is the predicted mental health risk. For example, abnormal emotional and behavioral patterns are detected by comparing them with past data.

[0607] Step 7:

[0608] Generates warnings and response messages

[0609] The server generates a warning message and a corresponding response message based on the risk prediction result. The input is the mental health risk prediction result, and the output is the generated warning message and corresponding response message. For example, a message such as "Your mental health risk is increasing. Take a rest or consult a specialist" may be generated.

[0610] Step 8:

[0611] User Notifications

[0612] The device notifies the user of the generated message. The notification is via a smartphone, smart glasses, or similar device. The input is the generated message, and the output is a notification displayed on the user's device. The user reviews the notification and receives information about their mental health condition and how to respond.

[0613] This processing step makes it possible to detect mental health risks early from the content of social media posts and notify users in real time.

[0614] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0615] The system of the present invention analyzes data posted on social networking sites to predict and warn of mental health risks, and in particular incorporates an emotion engine to provide a function for recognizing users' emotions in detail. Specific embodiments for implementing the present invention are described below.

[0616] 1. Data Collection

[0617] The server collects user posted data using the API of the SNS platform. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[0618] 2. Data Classification

[0619] The server categorizes the collected data into text, images, and videos, and sends the categorized data to the corresponding analysis module.

[0620] 3. Data Analysis

[0621] The server analyzes the data using various analysis modules.

[0622] The text data is analyzed using natural language processing (NLP) algorithms. Specifically, keyword extraction and sentiment analysis are performed to detect, for example, the keyword "fatigue" and its associated sentiment (fatigue).

[0623] The image data is then subjected to computer vision techniques for facial expression recognition, object recognition, and scene analysis, for example to detect neutral faces and dark backgrounds.

[0624] Video data is analyzed using a video analysis algorithm, which uses voice recognition and video analysis to perform multifaceted emotion analysis, such as detecting low voice tones and little movement.

[0625] 4. Emotion Recognition by Emotion Engine

[0626] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[0627] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions.

[0628] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions, to identify the emotional state in the image.

[0629] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc.

[0630] 5. Mental health risk prediction

[0631] The server integrates the emotional data recognized by the emotion engine and predicts mental health risks using a generative AI model, which compares the data with the user's past data to detect abnormal emotional and behavioral patterns.

[0632] 6. Generating warning and response messages

[0633] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0634] 7. User Notices

[0635] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[0636] Specific examples

[0637] Example 1: For user A

[0638] User A posts on social media, "I've been feeling really tired lately."

[0639] The server collects the posts and sends them to a text analysis module.

[0640] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0641] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0642] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[0643] A generative AI model integrates this data and predicts elevated mental health risks.

[0644] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[0645] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[0646] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[0647] The processing flow will be explained below.

[0648] Step 1:

[0649] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[0650] Step 2:

[0651] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[0652] Step 3:

[0653] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[0654] Step 4:

[0655] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[0656] Step 5:

[0657] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral expressions or dark backgrounds.

[0658] Step 6:

[0659] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[0660] Step 7:

[0661] Emotion recognition by emotion engine. The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[0662] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions, for example, recognizing negative emotions from the phrase "tired."

[0663] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions.

[0664] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc. For example, a low voice tone and little movement can indicate fatigue.

[0665] Step 8:

[0666] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[0667] Step 9:

[0668] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[0669] Step 10:

[0670] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[0671] Step 11:

[0672] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0673] Step 12:

[0674] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[0675] Step 13:

[0676] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[0677] Example 2

[0678] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0679] Conventional systems have had difficulty accurately predicting users' mental health risks from data posted on social media and generating appropriate warning and response messages. Furthermore, it is not easy to comprehensively analyze a variety of data (text, images, and videos) with explicit permission from the user. This makes it difficult to address users' mental health issues early, and there is a need for improved accuracy in risk prediction.

[0680] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0681] In this invention, the server includes means for acquiring posted data on SNS, means for classifying the posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for performing detailed emotion recognition on the analyzed text, image, and video data using an emotion engine, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages, and means for notifying the generated messages to the user. This makes it possible to accurately predict a user's mental health risks from the posted data on SNS and provide early and appropriate responses.

[0682] "SNS posting data" refers to content such as text, images, and videos posted by users on social networking services (SNS).

[0683] "Text" refers to data containing text information posted by users on SNS.

[0684] "Images" are data containing visual content posted by users on social networking sites, including photographs and illustrations.

[0685] "Video" refers to data containing content accompanied by motion and sound that users post on social networking sites.

[0686] "Natural language processing" is a technology that uses computers to analyze and understand human language, and includes text data analysis, keyword extraction, sentiment analysis, and more.

[0687] "Computer vision" is a technology that allows computers to analyze and understand visual information such as images and videos.

[0688] "Voice analysis" is a technology that analyzes voice data and performs functions such as converting it into text, analyzing emotions, and recognizing speech.

[0689] An "emotion engine" is a technology that comprehensively analyzes data such as text, images, and videos to recognize the user's emotional state in detail.

[0690] A "generative AI model" is an artificial intelligence model that uses machine learning technology to recognize specific patterns from input data and make predictions or generate results.

[0691] "Mental health risk" refers to the possibility that a user may have problems with their mental health, and is predicted based on abnormal emotions and behavioral patterns from analytical data.

[0692] A "warning message" is a message that notifies the user that their mental health risk is increasing.

[0693] The "response message" is a message that suggests specific ways to respond to a user whose mental health risk is increasing.

[0694] "Means of notification" refers to technology for transmitting the generated warning message and response message to the user, and includes dedicated apps, email, push notifications, etc.

[0695] The system of the present invention analyzes data posted on social media platforms to predict and warn of mental health risks. Specifically, the server uses the social media platform's API to collect user posted data (text, images, and videos), and classifies and analyzes this data to predict mental health risks. Furthermore, an emotion engine is incorporated to recognize users' emotions in detail, and appropriate warning and response messages are generated and sent.

[0696] Data collection

[0697] The server collects user posting data using the APIs of social networking platforms such as Twitter and Facebook. Multimodal data such as text, images, and videos are collected with the user's explicit permission. For example, if a user posts "I've been feeling tired lately," the text data and attached image and video data are also collected.

[0698] Data Classification

[0699] The server categorizes the collected posting data into three types: text, images, and videos. For example, a photo of a person not smiling or a video in a dark room posted together with a text message saying "I'm not feeling well" would be categorized separately. This ensures that data is properly sorted according to each analysis module.

[0700] Data analysis

[0701] The server analyzes the data using the following analysis modules:

[0702] The text data is then subjected to keyword extraction and sentiment analysis using NLP (natural language processing) algorithms, for example, to detect the keyword "fatigue" and the associated "feeling of fatigue."

[0703] The image data is then subjected to facial expression recognition and scene analysis using computer vision techniques, such as detecting neutral facial expressions and dark backgrounds as features.

[0704] Video data is analyzed using voice recognition and video analysis to detect voice tone and lack of movement, for example, low voice tone and lack of movement.

[0705] Emotion recognition by emotion engine

[0706] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[0707] From the text data, phrase analysis is performed to recognize positive, negative, and neutral emotions. For example, a post saying "I've been feeling tired lately" can be interpreted as a negative emotion.

[0708] Emotions are recognized from image data by analyzing facial expressions and scenes, for example, expressionless faces and dark scenes are detected.

[0709] Video data is analyzed for audio and video to detect low tones and little movement.

[0710] Mental health risk prediction

[0711] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risk. This model compares a user's past emotional data with their current data to detect abnormal emotional and behavioral patterns. For example, if a user's recent posts show an abnormally negative trend compared to past data, it predicts a high mental health risk.

[0712] Generates warnings and response messages

[0713] If the server predicts a high mental health risk, it generates a warning message and a corresponding message, such as "Your mental health is at increased risk. Take a rest or consult a specialist."

[0714] User Notifications

[0715] The device will notify the user of the warning message and corresponding message sent from the server via a dedicated app, email, or push notification. Users can check the notification, understand their mental health risk, and take necessary action.

[0716] Specific examples

[0717] Example 1: For user A

[0718] User A posts on social media, "I've been feeling really tired lately."

[0719] The server collects the posts and sends them to a text analysis module.

[0720] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0721] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0722] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[0723] A generative AI model integrates this data and predicts elevated mental health risks.

[0724] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[0725] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[0726] Example prompts for generative AI models

[0727] "Please tell me the system procedure for analyzing a user's emotional state from social media posts and predicting mental health risks."

[0728] "Please explain each processing step of a system that analyzes emotions based on user posts on social media and predicts mental health risks."

[0729] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks in users from posts on social media at an early stage and provide appropriate responses promptly.

[0730] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0731] Step 1: Data collection

[0732] The server uses the API of the SNS platform to collect user posted data. API authentication information (API key and secret) and user ID are required as input. The acquired data includes text, images, and videos. Specifically, the server sends an API request to the SNS platform and stores the user posted data in a temporary database.

[0733] Step 2: Data Classification

[0734] The server classifies the collected posted data into text, images, and videos. The input requires the multimodal data collected in step 1. The output is data classified into text data, image data, and video data. Specifically, the server checks the data format and distributes the data to the appropriate analysis module.

[0735] Step 3: Data analysis

[0736] The server analyzes data using various analysis modules. Text data, image data, and video data are required as input. Analysis results (keywords, emotional state, facial expression, voice tone, etc.) are obtained as output. The specific operation is as follows:

[0737] NLP algorithms are used to extract keywords and analyze sentiment from text data. For example, the keyword "tired" is extracted and the sentiment "fatigue" is determined.

[0738] Facial expression recognition is performed on image data using computer vision techniques. For example, neutral faces are detected.

[0739] Speech recognition and video analysis are performed on video data to detect voice tones and lack of movement. Example: Low voice tones and lack of movement are detected.

[0740] Step 4: Emotion Recognition with the Emotion Engine

[0741] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The analyzed data from step 3 is required as input. The output is a detailed emotional state (positive, negative, neutral). Specifically, the emotion engine performs phrase analysis, facial expression analysis, and voice analysis to recognize emotions from each media.

[0742] Step 5: Mental health risk prediction

[0743] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risks. Detailed emotional data and past emotional data are required as input. The output is a predicted mental health risk. Specifically, the system inputs emotional data into the generative AI model and performs risk assessment. In particular, it compares past data with current data to detect abnormal emotional and behavioral patterns.

[0744] Step 6: Generate warning and response messages

[0745] The server generates a warning message and a response message if a high mental health risk is predicted. The required input is the risk assessment result. The output is a warning message and a response message. Specifically, based on the risk assessment, the server generates a warning message such as "Take a rest or consult a specialist" and a message suggesting specific response methods such as "Talk to a friend" or "Consult a doctor."

[0746] Step 7: User Notification

[0747] The device notifies the user of the warning message and response message sent from the server. The generated message is required as input. The output is the completion of the notification to the user. The specific operation is to display the message via a dedicated app, email, or push notification. The user can then check the notification and take appropriate action.

[0748] (Application example 2)

[0749] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0750] In modern society, with the spread of self-driving vehicles, mental health issues among drivers are on the rise, and there is a need for a system that can monitor their mental state in real time and encourage appropriate responses. However, conventional technologies predict mental health risks solely from data posted on social media, making it difficult to reflect real-time conditions while driving. In addition, warning messages are not generated and notified quickly, making it difficult for drivers to recognize problems in a timely manner and take appropriate measures. These issues need to be resolved.

[0751] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring posted data on an SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and speech analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the user of the generated messages, means for analyzing data collected from sensors in the autonomous vehicle and evaluating the driver's mental state, and means for displaying a warning message using the vehicle's dashboard display system based on the evaluation results. This enables real-time monitoring of the driver's mental state while driving, and prompt warnings and response suggestions.

[0752] "SNS" is an abbreviation for social networking service, a platform for users to communicate and share information online.

[0753] "Posted data" refers to all information, including text, images, and videos, posted by users on SNS.

[0754] A "server" is a computer system that processes and manages information on a network.

[0755] "Natural language processing" refers to techniques and approaches that enable computers to understand and generate human language.

[0756] "Computer vision" is a technology that allows computers to analyze digital images and videos and understand visual information.

[0757] "Voice analysis" is a technology that analyzes voice data and extracts its content and characteristics.

[0758] "Mental health risk" refers to the possibility that an individual's mental health will deteriorate.

[0759] A "warning message" is a message that the system uses to notify the user of risks or situations that require caution.

[0760] A "response message" is a notification message that suggests specific actions or countermeasures for responding to a warning.

[0761] An "autonomous vehicle" is a vehicle that can be driven and navigated automatically without the need for human operation.

[0762] A "sensor" is a device that detects changes in the physical environment and acquires that information as electronic data.

[0763] A "dashboard" is the part of a vehicle that contains a display and interface through which the driver can view information.

[0764] This invention is a system that analyzes data posted on social media to predict and warn drivers about mental health risks, and is particularly applicable to self-driving vehicles. The system is configured as follows:

[0765] Hardware

[0766] Camera: Installed inside the vehicle to capture the driver's facial expressions and movements.

[0767] Microphone: Installed inside the vehicle to collect the driver's voice and audio.

[0768] Smart glasses: Worn by the driver, they track eye movements and gaze in real time.

[0769] Dashboard display: A display for showing warning and response messages.

[0770] software

[0771] SNS API: Retrieves post data from social media platforms.

[0772] Natural language processing (NLP) engine: Analyzes text data, extracts keywords, and performs sentiment analysis.

[0773] Computer vision tools: Analyze image data and perform facial expression and object recognition.

[0774] Video analysis algorithm: Analyzes video data and detects emotions from audio and video.

[0775] Emotion engine: Integrates various analytical data to recognize the driver's emotional state in detail.

[0776] Generative AI model: Predicting mental health risks based on integrated emotional data.

[0777] Notification system: Notifies the driver of warning and response messages via the dashboard display.

[0778] Program processing

[0779] The server collects driver posting data using SNS APIs. The collected data is categorized into text, images, and videos, and each data is sent to the corresponding analysis module. The NLP engine analyzes the text data to extract keywords and perform sentiment analysis. The computer vision tool analyzes the image data to recognize facial expressions and scenes. The video analysis algorithm analyzes the video data and extracts emotions from audio and video.

[0780] The analyzed data is integrated into an emotion engine to recognize the driver's emotional state in detail. The generative AI model uses this emotional data to predict mental health risks and immediately generates a warning message if the risk is high. For example, it might say, "Your mental health risk is increasing. Please take a break or consult a specialist." It also generates a response message suggesting specific ways to respond.

[0781] These messages are displayed on the dashboard display, allowing the driver to check their own mental health status in real time and take necessary action. This system makes it possible to detect mental health risks while driving early and provide appropriate responses quickly.

[0782] Specific examples

[0783] For example, if Driver B posts on social media while driving, "I'm tired right now, but I'm driving hard," the server collects this post, extracts the keyword "tired" using an NLP engine, and determines the state as "fatigue" using sentiment analysis. Furthermore, the vehicle's camera detects expressionless faces, and the microphone recognizes monotone voices. This data is integrated, and the sentiment engine precisely recognizes the emotions of "fatigue" and "expressionless," and the generative AI model predicts a high mental health risk. The system generates a warning message saying, "Your mental health risk is increasing. You should take a break," and displays this on the dashboard display. Driver B sees this message and takes action to rest in a safe place.

[0784] Prompt Sentence Examples

[0785] Input prompt: "Based on recent posts, predict risks and generate a warning message."

[0786] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0787] Step 1:

[0788] The server collects driver posting data using SNS API. The collected data is in the form of text, images, and videos, and is executed with the user's explicit permission. The SNS posting data is used as input, and it is output in the form of classified data.

[0789] Step 2:

[0790] The server classifies the collected data into text, images, and videos. The input here is the various posted data collected in step 1, and the classified data format (text, images, videos) is output. This classification is done automatically.

[0791] Step 3:

[0792] The server analyzes the classified data. Specifically, text data is analyzed using a natural language processing (NLP) engine to extract keywords and perform sentiment analysis. Image data is analyzed using computer vision tools to recognize facial expressions and scenes. Video data is analyzed using video analysis algorithms to extract emotions from audio and video. The input is the data classified in step 2, and the analyzed emotional data is the output.

[0793] Step 4:

[0794] The server integrates the analyzed data and performs detailed emotion recognition using an emotion engine. Here, the analysis results of each piece of text, image, and video data are input, and the integrated emotion data is output. The emotion engine recognizes the emotional state of each piece of data in detail and provides integrated emotion information.

[0795] Step 5:

[0796] The server uses a generative AI model based on the integrated emotion data to predict mental health risks. The input is the integrated emotion data, and the output is the predicted mental health risk. This prediction is made by comparing it with past data to detect abnormal emotion and behavior patterns.

[0797] Step 6:

[0798] If the server predicts a high mental health risk, it immediately generates a warning message and a corresponding message. The input here is the predicted mental health risk, and the message output is something like "Your mental health risk is increasing. Take a rest or consult a specialist."

[0799] Step 7:

[0800] The terminal notifies the driver of the generated message. Specifically, it displays warning messages and response messages in real time on the dashboard display so that the driver can check them. The input is the message generated in step 6, and the output is the displayed message. The driver can check this message and take safety measures.

[0801] For example, if Driver B posts on social media, "I'm tired right now, but I'm driving hard," the camera will detect an expressionless face and the microphone will detect a monotone voice. These data will be integrated to recognize the emotions of "fatigue" and "expressionless," and the generative AI model will predict an increase in mental health risk and display an appropriate warning message.

[0802] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0803] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0804] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0805] [Third embodiment]

[0806] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0807] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0808] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0809] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0810] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0811] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0812] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0813] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0814] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0815] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0816] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0817] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0818] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[0819] 1. Data Collection

[0820] The server collects user posted data using the API of the SNS platform. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[0821] 2. Data Classification

[0822] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[0823] 3. Data Analysis

[0824] The server analyzes the data using various analysis modules.

[0825] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[0826] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[0827] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[0828] 4. Mental health risk prediction

[0829] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[0830] 5. Generating warning and response messages

[0831] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0832] 6. User Notices

[0833] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[0834] Specific examples

[0835] Example 1: For user A

[0836] User A posts on social media, "I've been feeling really tired lately."

[0837] The server collects the posts and sends them to a text analysis module.

[0838] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0839] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0840] A generative AI model integrates this data and predicts elevated mental health risks.

[0841] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[0842] The terminal notifies this message to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[0843] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[0844] The processing flow will be explained below.

[0845] Step 1:

[0846] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[0847] Step 2:

[0848] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[0849] Step 3:

[0850] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[0851] Step 4:

[0852] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[0853] Step 5:

[0854] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral expressions or dark backgrounds.

[0855] Step 6:

[0856] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[0857] Step 7:

[0858] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[0859] Step 8:

[0860] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[0861] Step 9:

[0862] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[0863] Step 10:

[0864] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0865] Step 11:

[0866] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[0867] Step 12:

[0868] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[0869] Example 1

[0870] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0871] In recent years, with the spread of social networking sites, it has become increasingly difficult for users to recognize their own mental health risks. In particular, in today's world where posting on social networking sites has become a part of everyday life, technology that can analyze this posting data and detect users' mental health status early is crucial. However, existing systems lack the functionality to perform multifaceted analysis of posting data, making it difficult to accurately predict mental health risks and provide users with appropriate warnings and response messages.

[0872] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0873] In this invention, the server includes means for acquiring posted data on SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and video analysis, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages based on the prediction results, and means for notifying the generated messages to the user. This enables multifaceted analysis of the user's mental health status from SNS posted data, and enables highly accurate risk prediction and generation of warning and response messages.

[0874] "Means for obtaining data posted on SNS" refers to a function that uses the API of a social networking service (SNS) platform to collect data such as text, images, and videos posted with the user's permission.

[0875] "Means for classifying acquired posted data into text, images, and videos" is a function that detects the format of collected posted data and sorts it into the appropriate module as text data, image data, or video data.

[0876] "Means for analyzing classified data using natural language processing, computer vision, and video analysis" refers to functions for analyzing text data using natural language processing technology (NLP), analyzing image data using computer vision technology, and analyzing video data using video analysis technology.

[0877] "Means for integrating analyzed data and using a generative AI model to predict mental health risks" refers to a function that integrates analyzed text, image, and video data and uses a generative AI model to predict a user's mental health risks.

[0878] The "means for generating a warning message and a response message based on the prediction result" is a function for generating a warning message for the user and a message suggesting specific response methods based on the mental health risk prediction result.

[0879] The "means for notifying the user of the generated message" is a function for sending the generated warning message and corresponding message to the user via a dedicated app, email, or push notification.

[0880] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[0881] Hardware and software used

[0882] Hardware:

[0883] Server: High-performance server (e.g. AWS EC2, Google Cloud Compute Engine)

[0884] Device: Smartphone (e.g. iPhone, Android device)

[0885] software:

[0886] Natural Language Processing algorithms: BERT and GPT-based models

[0887] Computer Vision Technologies: OpenCV, TensorFlow, Keras

[0888] Video analysis algorithms: OpenPose, YOLOv3

[0889] Generative AI models: GPT-3 and the latest generative AI models

[0890] Data collection

[0891] The server collects user post data through the API of the social media platform. Specifically, the server periodically calls the API to retrieve user posts (text, images, videos). This also includes a step to obtain the user's explicit permission. For example, the server uses the Twitter API to call "GET statuses / user_timeline" to collect the latest posts from a specified user.

[0892] Data Classification

[0893] The server classifies the collected data into three modules: text, image, and video. Specifically, it checks the data format and sends the data to the appropriate analysis module. For example, text data is sent to the NLP (natural language processing) module, image data is sent to the computer vision module, and video data is sent to the video analysis module.

[0894] Data analysis

[0895] The server analyzes the data in each module. The specific operations are as follows:

[0896] Text data: The NLP module performs keyword extraction and sentiment analysis using BERT and GPT-based models. For example, it extracts the keyword "tired" and determines the sentiment as "fatigue."

[0897] Image data: The computer vision module uses OpenCV and TensorFlow to perform facial expression recognition, object recognition, and scene analysis, for example, analyzing whether a face in an image has a neutral expression.

[0898] Video data: The video analysis module uses speech recognition and video analysis algorithms (e.g., OpenPose, YOLOv3) to perform multifaceted sentiment analysis, such as analyzing changes in voice tone and facial expressions.

[0899] Mental health risk prediction

[0900] The server integrates the analysis results and uses a generative AI model to predict mental health risks. Specifically, it compares the user's past data with the current analysis results to detect abnormal emotional and behavioral patterns. For example, if the user's posted text frequently contains the word "tired," this will be added up as a risk.

[0901] Generates warnings and response messages

[0902] If the server predicts a high mental health risk, it immediately generates a warning message. Specifically, it generates messages such as "Consult a specialist" or "Take a rest" depending on the risk level. The generative AI model also suggests specific actions that the user can take. For example, it generates a message that reads, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist."

[0903] User Notifications

[0904] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and are selected based on the user's settings. For example, the push notification function of a smartphone app can be used to display a message saying "Consult an expert." However, email notification is also possible.

[0905] Specific examples

[0906] Example 1: For user A

[0907] User A posts on social media, "I've been feeling really tired lately."

[0908] The server collects these posts using the SNS API and sends them to the text analysis module.

[0909] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[0910] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[0911] A generative AI model integrates this data to predict elevated mental health risks.

[0912] The server displays a warning message saying, "Your mental health is at increased risk. Take a rest or consult a professional," along with a suggestion for specific action: "Start practicing mindfulness to reduce stress."

[0913] The device sends this message as a push notification to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[0914] Prompt Sentence Examples

[0915] "Predict the mental health risk of users who recently posted on social media that they've been feeling very tired lately."

[0916] "Identify the mental health risk from the following post: 'Image: Expressionless face, Text: I've been feeling tired lately' and generate an appropriate warning message."

[0917] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks in users early from SNS posting data and respond promptly.

[0918] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0919] Step 1: Data collection

[0920] The server collects user post data using the API of the SNS platform. Specifically, the server periodically makes API calls to retrieve user posts (text, images, videos) using endpoints such as "GET statuses / user_timeline". The input is the post data obtained as a result of the SNS API call, and the output is the retrieved raw data.

[0921] Step 2: Data Classification

[0922] The server classifies the collected posted data into three types: text, image, and video. The server checks the data format and distributes text data to the NLP module, image data to the computer vision module, and video data to the video analysis module. The input is the collected raw data, and the output is data classified by format.

[0923] Step 3: Text data analysis

[0924] The server processes the text data in a text analysis module. Specifically, it uses BERT or GPT-based natural language processing (NLP) models to extract keywords and perform sentiment analysis. For example, from the text posted by a user saying "I've been feeling tired lately," it extracts the keyword "tired" and recognizes the emotion as "fatigue." The input is text data, and the output is the analyzed text data and extracted sentiment information.

[0925] Step 4: Image data analysis

[0926] The server processes image data using an image analysis module. Specifically, it uses OpenCV and TensorFlow to recognize facial expressions in images and determine their emotions. For example, it performs expressionless face detection and background scene analysis. The input is image data, and the output is analyzed image data and emotional information.

[0927] Step 5: Video data analysis

[0928] The server processes the video data using a video analysis module. Specifically, it uses OpenPose and YOLOv3 to analyze the audio and video in the video and detect emotional changes. For example, it analyzes changes in voice tone and facial expressions. The input is the video data, and the output is the analyzed video data and emotional information.

[0929] Step 6: Data integration and mental health risk prediction

[0930] The server integrates the analysis results obtained from text, images, and videos and uses a generative AI model to predict mental health risks. Specifically, it centralizes this data and compares it with the user's past data to detect abnormal patterns. The input is the analyzed text, image, and video data, and the output is the predicted mental health risk.

[0931] Step 7: Generate warning and response messages

[0932] The server generates a warning message based on the risk prediction result. For example, it generates a message such as "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist," or a message suggesting specific measures to take, such as "Take a break" or "Talk to a friend." The input is the risk prediction result, and the output is the generated warning message and response message.

[0933] Step 8: User Notification

[0934] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and notifications are sent in the appropriate way depending on the user's settings. For example, the push notification function of a smartphone app can be used to display the generated warning message. The input is the generated warning message and the corresponding message, and the output is the notification sent to the user.

[0935] (Application example 1)

[0936] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0937] In recent years, the spread of social media has led to many people sharing information on a daily basis. However, at the same time, there has been an increase in the number of mental health risks hidden in posts on social media, and there is a need to detect these risks early and take appropriate measures. In addition, there is a problem that it is difficult to respond in real time due to the lack of technology that can instantly notify users of risks using devices such as smartphones or smart glasses.

[0938] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0939] In this invention, the server includes means for acquiring posted data on the SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the generated messages to the user, and means for notifying the user of the warning messages via a smartphone, smart glasses, or similar device. This enables early detection of mental health risks from the content of posts on the SNS and notification to the user in real time.

[0940] "SNS" is an online platform that allows users to share their information and opinions over the Internet.

[0941] "Posted data" refers to digital content such as text, images, and videos posted by users on SNS.

[0942] "Text classification" is the process of analyzing acquired post data based on the content of the strings and sorting them into specific categories.

[0943] "Image classification" is a technology that analyzes images in acquired posting data and classifies them into specific categories based on their content.

[0944] "Video classification" is a technology that analyzes videos in the acquired posting data and classifies them based on their content and characteristics.

[0945] "Natural language processing" is a general term for algorithms and technologies that allow computers to understand and analyze human language.

[0946] "Computer vision" refers to the technology and algorithms used to analyze and understand image and video data.

[0947] "Voice analysis" is a technology that analyzes voice data in digital format and recognizes its content and characteristics.

[0948] "Mental health risks" refer to conditions or symptoms that may have a negative impact on mental health.

[0949] A "warning message" is alert information that notifies the user of a crisis or risk.

[0950] The "response message" is information that suggests specific actions or methods of response to the user who receives the warning message.

[0951] A "generative AI model" is a technology that uses an artificial intelligence model that learns from existing data and experience to predict the characteristics and risks of new data.

[0952] "Notification means" refers to the method or technique for notifying the user of the generated message.

[0953] A "smartphone" is a mobile phone device with advanced computing power and connectivity.

[0954] "Smart glasses" are wearable devices that can integrate and display real and digital information.

[0955] "Device" is a general term for electronic devices that have specific functions or roles.

[0956] The system for implementing the present invention includes a wide range of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. A specific method for implementing this system is described below.

[0957] 1. Data Collection

[0958] The server uses the API of the SNS platform to collect user posted data. This requires the user's explicit permission. The collected data is multimodal data such as text, images, and videos.

[0959] 2. Data Classification

[0960] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[0961] 3. Data Analysis

[0962] The server analyzes the data using various analysis modules.

[0963] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[0964] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[0965] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[0966] 4. Mental health risk prediction

[0967] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[0968] 5. Generating warning and response messages

[0969] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[0970] 6. User Notices

[0971] The device notifies the user of the generated message via a dedicated app, email, or push notification. Additionally, notifications can be sent via smartphones, smart glasses, or similar devices. The user can then review the notifications to understand their mental health status and any necessary actions.

[0972] Specific examples

[0973] Example 1: For user A

[0974] User A posts on social media, "I've been feeling really tired lately." The server collects this post and sends it to a text analysis module. An NLP algorithm extracts the keyword "tired," and sentiment analysis determines the feeling as "fatigue." Computer vision analysis detects that the posted image contains an inconspicuous background and an expressionless face. A generative AI model integrates this data and predicts a high mental health risk. The server generates a warning message for User A saying, "Your mental health risk is increasing. Please take a rest or consult a specialist." The device notifies User A of this message. User A checks the notification and takes appropriate action, such as consulting a specialist.

[0975] Example prompts to input to the generative AI model

[0976] "Analyze the following text and image posted by a user on social media and predict their mental health risk. Text: "I'm so sad", Image: <Image URL>"

[0977] This system makes it possible to detect mental health risks from social media posts early on and provide prompt responses.

[0978] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0979] Step 1:

[0980] Data collection

[0981] The server collects user posted data using the API of the SNS platform. First, with the user's explicit permission, it retrieves text, image, and video data from the SNS using a dedicated app and API key. The input is the user's SNS account information and API key, and the output is the collected SNS posted data.

[0982] Step 2:

[0983] Data Classification

[0984] The server classifies the collected data. The acquired text data is sent to the text analysis module, the image data is sent to the image analysis module, and the video data is sent to the video analysis module. The input is the collected SNS post data, and the output is classified text, image, and video data.

[0985] Step 3:

[0986] Text Data Analysis

[0987] The server analyzes the text data using natural language processing (NLP) algorithms. Specifically, it extracts keywords and performs sentiment analysis. The input is text data, and the output is analyzed emotional information and keywords. For example, the server can detect the emotion "sad" from the posted text "very sad."

[0988] Step 4:

[0989] Image data analysis

[0990] The server analyzes image data using computer vision technology. Specifically, it performs facial expression recognition, object recognition, and scene analysis. The input is image data, and the output is analyzed facial expression information and scene information. For example, neutral faces and dark backgrounds are detected.

[0991] Step 5:

[0992] Video Data Analysis

[0993] The server uses a video analysis algorithm to perform multifaceted emotion analysis using voice recognition and video analysis. The input is video data, and the output is analyzed emotional information and behavioral patterns. For example, if the user's voice tone is low and the video is dark, a report will be generated.

[0994] Step 6:

[0995] Data integration and risk prediction

[0996] The server integrates the analyzed text, image, and video data and uses a generative AI model to predict mental health risks. The input is multiple analyzed data (emotional information, facial expression information, and behavioral patterns), and the output is the predicted mental health risk. For example, abnormal emotional and behavioral patterns are detected by comparing them with past data.

[0997] Step 7:

[0998] Generates warnings and response messages

[0999] The server generates a warning message and a corresponding response message based on the risk prediction result. The input is the mental health risk prediction result, and the output is the generated warning message and corresponding response message. For example, a message such as "Your mental health risk is increasing. Take a rest or consult a specialist" may be generated.

[1000] Step 8:

[1001] User Notifications

[1002] The device notifies the user of the generated message. The notification is via a smartphone, smart glasses, or similar device. The input is the generated message, and the output is a notification displayed on the user's device. The user reviews the notification and receives information about their mental health condition and how to respond.

[1003] This processing step makes it possible to detect mental health risks early from the content of social media posts and notify users in real time.

[1004] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1005] The system of the present invention analyzes data posted on social networking sites to predict and warn of mental health risks, and in particular incorporates an emotion engine to provide a function for recognizing users' emotions in detail. Specific embodiments for implementing the present invention are described below.

[1006] 1. Data Collection

[1007] The server collects user posted data using the API of the SNS platform. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[1008] 2. Data Classification

[1009] The server categorizes the collected data into text, images, and videos, and sends the categorized data to the corresponding analysis module.

[1010] 3. Data Analysis

[1011] The server analyzes the data using various analysis modules.

[1012] The text data is analyzed using natural language processing (NLP) algorithms. Specifically, keyword extraction and sentiment analysis are performed to detect, for example, the keyword "fatigue" and its associated sentiment (fatigue).

[1013] The image data is then subjected to computer vision techniques for facial expression recognition, object recognition, and scene analysis, for example to detect neutral faces and dark backgrounds.

[1014] Video data is analyzed using a video analysis algorithm, which uses voice recognition and video analysis to perform multifaceted emotion analysis, such as detecting low voice tones and little movement.

[1015] 4. Emotion Recognition by Emotion Engine

[1016] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[1017] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions.

[1018] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions, to identify the emotional state in the image.

[1019] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc.

[1020] 5. Mental health risk prediction

[1021] The server integrates the emotional data recognized by the emotion engine and predicts mental health risks using a generative AI model, which compares the data with the user's past data to detect abnormal emotional and behavioral patterns.

[1022] 6. Generating warning and response messages

[1023] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[1024] 7. User Notices

[1025] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[1026] Specific examples

[1027] Example 1: For user A

[1028] User A posts on social media, "I've been feeling really tired lately."

[1029] The server collects the posts and sends them to a text analysis module.

[1030] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[1031] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[1032] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[1033] A generative AI model integrates this data and predicts elevated mental health risks.

[1034] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[1035] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[1036] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[1037] The processing flow will be explained below.

[1038] Step 1:

[1039] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[1040] Step 2:

[1041] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[1042] Step 3:

[1043] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[1044] Step 4:

[1045] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[1046] Step 5:

[1047] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral expressions or dark backgrounds.

[1048] Step 6:

[1049] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[1050] Step 7:

[1051] Emotion recognition by emotion engine. The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[1052] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions, for example, recognizing negative emotions from the phrase "tired."

[1053] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions.

[1054] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc. For example, a low voice tone and little movement can indicate fatigue.

[1055] Step 8:

[1056] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[1057] Step 9:

[1058] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[1059] Step 10:

[1060] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[1061] Step 11:

[1062] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[1063] Step 12:

[1064] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[1065] Step 13:

[1066] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[1067] Example 2

[1068] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1069] Conventional systems have had difficulty accurately predicting users' mental health risks from data posted on social media and generating appropriate warning and response messages. Furthermore, it is not easy to comprehensively analyze a variety of data (text, images, and videos) with explicit permission from the user. This makes it difficult to address users' mental health issues early, and there is a need for improved accuracy in risk prediction.

[1070] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1071] In this invention, the server includes means for acquiring posted data on SNS, means for classifying the posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for performing detailed emotion recognition on the analyzed text, image, and video data using an emotion engine, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages, and means for notifying the generated messages to the user. This makes it possible to accurately predict a user's mental health risks from the posted data on SNS and provide early and appropriate responses.

[1072] "SNS posting data" refers to content such as text, images, and videos posted by users on social networking services (SNS).

[1073] "Text" refers to data containing text information posted by users on SNS.

[1074] "Images" are data containing visual content posted by users on social networking sites, including photographs and illustrations.

[1075] "Video" refers to data containing content accompanied by motion and sound that users post on social networking sites.

[1076] "Natural language processing" is a technology that uses computers to analyze and understand human language, and includes text data analysis, keyword extraction, sentiment analysis, and more.

[1077] "Computer vision" is a technology that allows computers to analyze and understand visual information such as images and videos.

[1078] "Voice analysis" is a technology that analyzes voice data and performs functions such as converting it into text, analyzing emotions, and recognizing speech.

[1079] An "emotion engine" is a technology that comprehensively analyzes data such as text, images, and videos to recognize the user's emotional state in detail.

[1080] A "generative AI model" is an artificial intelligence model that uses machine learning technology to recognize specific patterns from input data and make predictions or generate results.

[1081] "Mental health risk" refers to the possibility that a user may have problems with their mental health, and is predicted based on abnormal emotions and behavioral patterns from analytical data.

[1082] A "warning message" is a message that notifies the user that their mental health risk is increasing.

[1083] The "response message" is a message that suggests specific ways to respond to a user whose mental health risk is increasing.

[1084] "Means of notification" refers to technology for transmitting the generated warning message and response message to the user, and includes dedicated apps, email, push notifications, etc.

[1085] The system of the present invention analyzes data posted on social media platforms to predict and warn of mental health risks. Specifically, the server uses the social media platform's API to collect user posted data (text, images, and videos), and classifies and analyzes this data to predict mental health risks. Furthermore, an emotion engine is incorporated to recognize users' emotions in detail, and appropriate warning and response messages are generated and sent.

[1086] Data collection

[1087] The server collects user posting data using the APIs of social networking platforms such as Twitter and Facebook. Multimodal data such as text, images, and videos are collected with the user's explicit permission. For example, if a user posts "I've been feeling tired lately," the text data and attached image and video data are also collected.

[1088] Data Classification

[1089] The server categorizes the collected posting data into three types: text, images, and videos. For example, a photo of a person not smiling or a video in a dark room posted together with a text message saying "I'm not feeling well" would be categorized separately. This ensures that data is properly sorted according to each analysis module.

[1090] Data analysis

[1091] The server analyzes the data using the following analysis modules:

[1092] The text data is then subjected to keyword extraction and sentiment analysis using NLP (natural language processing) algorithms, for example, to detect the keyword "fatigue" and the associated "feeling of fatigue."

[1093] The image data is then subjected to facial expression recognition and scene analysis using computer vision techniques, such as detecting neutral facial expressions and dark backgrounds as features.

[1094] Video data is analyzed using voice recognition and video analysis to detect voice tone and lack of movement, for example, low voice tone and lack of movement.

[1095] Emotion recognition by emotion engine

[1096] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[1097] From the text data, phrase analysis is performed to recognize positive, negative, and neutral emotions. For example, a post saying "I've been feeling tired lately" can be interpreted as a negative emotion.

[1098] Emotions are recognized from image data by analyzing facial expressions and scenes, for example, expressionless faces and dark scenes are detected.

[1099] Video data is analyzed for audio and video to detect low tones and little movement.

[1100] Mental health risk prediction

[1101] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risk. This model compares a user's past emotional data with their current data to detect abnormal emotional and behavioral patterns. For example, if a user's recent posts show an abnormally negative trend compared to past data, it predicts a high mental health risk.

[1102] Generates warnings and response messages

[1103] If the server predicts a high mental health risk, it generates a warning message and a corresponding message, such as "Your mental health is at increased risk. Take a rest or consult a specialist."

[1104] User Notifications

[1105] The device will notify the user of the warning message and corresponding message sent from the server via a dedicated app, email, or push notification. Users can check the notification, understand their mental health risk, and take necessary action.

[1106] Specific examples

[1107] Example 1: For user A

[1108] User A posts on social media, "I've been feeling really tired lately."

[1109] The server collects the posts and sends them to a text analysis module.

[1110] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[1111] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[1112] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[1113] A generative AI model integrates this data and predicts elevated mental health risks.

[1114] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[1115] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[1116] Example prompts for generative AI models

[1117] "Please tell me the system procedure for analyzing a user's emotional state from social media posts and predicting mental health risks."

[1118] "Please explain each processing step of a system that analyzes emotions based on user posts on social media and predicts mental health risks."

[1119] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks in users from posts on social media at an early stage and provide appropriate responses promptly.

[1120] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1121] Step 1: Data collection

[1122] The server uses the API of the SNS platform to collect user posted data. API authentication information (API key and secret) and user ID are required as input. The acquired data includes text, images, and videos. Specifically, the server sends an API request to the SNS platform and stores the user posted data in a temporary database.

[1123] Step 2: Data Classification

[1124] The server classifies the collected posted data into text, images, and videos. The input requires the multimodal data collected in step 1. The output is data classified into text data, image data, and video data. Specifically, the server checks the data format and distributes the data to the appropriate analysis module.

[1125] Step 3: Data analysis

[1126] The server analyzes data using various analysis modules. Text data, image data, and video data are required as input. Analysis results (keywords, emotional state, facial expression, voice tone, etc.) are obtained as output. The specific operation is as follows:

[1127] NLP algorithms are used to extract keywords and analyze sentiment from text data. For example, the keyword "tired" is extracted and the sentiment "fatigue" is determined.

[1128] Facial expression recognition is performed on image data using computer vision techniques. For example, neutral faces are detected.

[1129] Speech recognition and video analysis are performed on video data to detect voice tones and lack of movement. Example: Low voice tones and lack of movement are detected.

[1130] Step 4: Emotion Recognition with the Emotion Engine

[1131] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The analyzed data from step 3 is required as input. The output is a detailed emotional state (positive, negative, neutral). Specifically, the emotion engine performs phrase analysis, facial expression analysis, and voice analysis to recognize emotions from each media.

[1132] Step 5: Mental health risk prediction

[1133] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risks. Detailed emotional data and past emotional data are required as input. The output is a predicted mental health risk. Specifically, the system inputs emotional data into the generative AI model and performs risk assessment. In particular, it compares past data with current data to detect abnormal emotional and behavioral patterns.

[1134] Step 6: Generate warning and response messages

[1135] The server generates a warning message and a response message if a high mental health risk is predicted. The required input is the risk assessment result. The output is a warning message and a response message. Specifically, based on the risk assessment, the server generates a warning message such as "Take a rest or consult a specialist" and a message suggesting specific response methods such as "Talk to a friend" or "Consult a doctor."

[1136] Step 7: User Notification

[1137] The device notifies the user of the warning message and response message sent from the server. The generated message is required as input. The output is the completion of the notification to the user. The specific operation is to display the message via a dedicated app, email, or push notification. The user can then check the notification and take appropriate action.

[1138] (Application example 2)

[1139] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1140] In modern society, with the spread of self-driving vehicles, mental health issues among drivers are on the rise, and there is a need for a system that can monitor their mental state in real time and encourage appropriate responses. However, conventional technologies predict mental health risks solely from data posted on social media, making it difficult to reflect real-time conditions while driving. In addition, warning messages are not generated and notified quickly, making it difficult for drivers to recognize problems in a timely manner and take appropriate measures. These issues need to be resolved.

[1141] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring posted data on an SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and speech analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the user of the generated messages, means for analyzing data collected from sensors in the autonomous vehicle and evaluating the driver's mental state, and means for displaying a warning message using the vehicle's dashboard display system based on the evaluation results. This enables real-time monitoring of the driver's mental state while driving, and prompt warnings and response suggestions.

[1142] "SNS" is an abbreviation for social networking service, a platform for users to communicate and share information online.

[1143] "Posted data" refers to all information, including text, images, and videos, posted by users on SNS.

[1144] A "server" is a computer system that processes and manages information on a network.

[1145] "Natural language processing" refers to techniques and approaches that enable computers to understand and generate human language.

[1146] "Computer vision" is a technology that allows computers to analyze digital images and videos and understand visual information.

[1147] "Voice analysis" is a technology that analyzes voice data and extracts its content and characteristics.

[1148] "Mental health risk" refers to the possibility that an individual's mental health will deteriorate.

[1149] A "warning message" is a message that the system uses to notify the user of risks or situations that require caution.

[1150] A "response message" is a notification message that suggests specific actions or countermeasures for responding to a warning.

[1151] An "autonomous vehicle" is a vehicle that can be driven and navigated automatically without the need for human operation.

[1152] A "sensor" is a device that detects changes in the physical environment and acquires that information as electronic data.

[1153] A "dashboard" is the part of a vehicle that contains a display and interface through which the driver can view information.

[1154] This invention is a system that analyzes data posted on social media to predict and warn drivers about mental health risks, and is particularly applicable to self-driving vehicles. The system is configured as follows:

[1155] Hardware

[1156] Camera: Installed inside the vehicle to capture the driver's facial expressions and movements.

[1157] Microphone: Installed inside the vehicle to collect the driver's voice and audio.

[1158] Smart glasses: Worn by the driver, they track eye movements and gaze in real time.

[1159] Dashboard display: A display for showing warning and response messages.

[1160] software

[1161] SNS API: Retrieves post data from social media platforms.

[1162] Natural language processing (NLP) engine: Analyzes text data, extracts keywords, and performs sentiment analysis.

[1163] Computer vision tools: Analyze image data and perform facial expression and object recognition.

[1164] Video analysis algorithm: Analyzes video data and detects emotions from audio and video.

[1165] Emotion engine: Integrates various analytical data to recognize the driver's emotional state in detail.

[1166] Generative AI model: Predicting mental health risks based on integrated emotional data.

[1167] Notification system: Notifies the driver of warning and response messages via the dashboard display.

[1168] Program processing

[1169] The server collects driver posting data using SNS APIs. The collected data is categorized into text, images, and videos, and each data is sent to the corresponding analysis module. The NLP engine analyzes the text data to extract keywords and perform sentiment analysis. The computer vision tool analyzes the image data to recognize facial expressions and scenes. The video analysis algorithm analyzes the video data and extracts emotions from audio and video.

[1170] The analyzed data is integrated into an emotion engine to recognize the driver's emotional state in detail. The generative AI model uses this emotional data to predict mental health risks and immediately generates a warning message if the risk is high. For example, it might say, "Your mental health risk is increasing. Please take a break or consult a specialist." It also generates a response message suggesting specific ways to respond.

[1171] These messages are displayed on the dashboard display, allowing the driver to check their own mental health status in real time and take necessary action. This system makes it possible to detect mental health risks while driving early and provide appropriate responses quickly.

[1172] Specific examples

[1173] For example, if Driver B posts on social media while driving, "I'm tired right now, but I'm driving hard," the server collects this post, extracts the keyword "tired" using an NLP engine, and determines the state as "fatigue" using sentiment analysis. Furthermore, the vehicle's camera detects expressionless faces, and the microphone recognizes monotone voices. This data is integrated, and the sentiment engine precisely recognizes the emotions of "fatigue" and "expressionless," and the generative AI model predicts a high mental health risk. The system generates a warning message saying, "Your mental health risk is increasing. You should take a break," and displays this on the dashboard display. Driver B sees this message and takes action to rest in a safe place.

[1174] Prompt Sentence Examples

[1175] Input prompt: "Based on recent posts, predict risks and generate a warning message."

[1176] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1177] Step 1:

[1178] The server collects driver posting data using SNS API. The collected data is in the form of text, images, and videos, and is executed with the user's explicit permission. The SNS posting data is used as input, and it is output in the form of classified data.

[1179] Step 2:

[1180] The server classifies the collected data into text, images, and videos. The input here is the various posted data collected in step 1, and the classified data format (text, images, videos) is output. This classification is done automatically.

[1181] Step 3:

[1182] The server analyzes the classified data. Specifically, text data is analyzed using a natural language processing (NLP) engine to extract keywords and perform sentiment analysis. Image data is analyzed using computer vision tools to recognize facial expressions and scenes. Video data is analyzed using video analysis algorithms to extract emotions from audio and video. The input is the data classified in step 2, and the analyzed emotional data is the output.

[1183] Step 4:

[1184] The server integrates the analyzed data and performs detailed emotion recognition using an emotion engine. Here, the analysis results of each piece of text, image, and video data are input, and integrated emotion data is output. The emotion engine recognizes the emotional state of each piece of data in detail and provides integrated emotion information.

[1185] Step 5:

[1186] The server uses a generative AI model based on the integrated emotion data to predict mental health risks. The input is the integrated emotion data, and the output is the predicted mental health risk. This prediction is made by comparing it with past data to detect abnormal emotion and behavior patterns.

[1187] Step 6:

[1188] If the server predicts a high mental health risk, it immediately generates a warning message and a corresponding message. The input here is the predicted mental health risk, and the message output is something like "Your mental health risk is increasing. Take a rest or consult a specialist."

[1189] Step 7:

[1190] The terminal notifies the driver of the generated message. Specifically, it displays warning messages and response messages in real time on the dashboard display so that the driver can check them. The input is the message generated in step 6, and the output is the displayed message. The driver can check this message and take safety measures.

[1191] For example, if Driver B posts on social media, "I'm tired right now, but I'm driving hard," the camera will detect an expressionless face and the microphone will detect a monotone voice. These data will be integrated to recognize the emotions of "fatigue" and "expressionless," and the generative AI model will predict an increase in mental health risk and display an appropriate warning message.

[1192] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1193] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1194] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1195] [Fourth embodiment]

[1196] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1197] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1198] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1199] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1200] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1201] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1202] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1203] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1204] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1205] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1206] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1207] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1208] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1209] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[1210] 1. Data Collection

[1211] The server collects user posted data using the API of the SNS platform. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[1212] 2. Data Classification

[1213] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[1214] 3. Data Analysis

[1215] The server analyzes the data using various analysis modules.

[1216] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[1217] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[1218] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[1219] 4. Mental health risk prediction

[1220] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[1221] 5. Generating warning and response messages

[1222] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[1223] 6. User Notices

[1224] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[1225] Specific examples

[1226] Example 1: For user A

[1227] User A posts on social media, "I've been feeling really tired lately."

[1228] The server collects the posts and sends them to a text analysis module.

[1229] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[1230] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[1231] A generative AI model integrates this data and predicts elevated mental health risks.

[1232] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[1233] The terminal notifies this message to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[1234] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[1235] The processing flow will be explained below.

[1236] Step 1:

[1237] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[1238] Step 2:

[1239] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[1240] Step 3:

[1241] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[1242] Step 4:

[1243] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[1244] Step 5:

[1245] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral expressions or dark backgrounds.

[1246] Step 6:

[1247] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[1248] Step 7:

[1249] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[1250] Step 8:

[1251] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[1252] Step 9:

[1253] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[1254] Step 10:

[1255] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[1256] Step 11:

[1257] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[1258] Step 12:

[1259] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[1260] Example 1

[1261] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1262] In recent years, with the spread of social networking sites, it has become increasingly difficult for users to recognize their own mental health risks. In particular, in today's world where posting on social networking sites has become a part of everyday life, technology that can analyze this posting data and detect users' mental health status early is crucial. However, existing systems lack the functionality to perform multifaceted analysis of posting data, making it difficult to accurately predict mental health risks and provide users with appropriate warnings and response messages.

[1263] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1264] In this invention, the server includes means for acquiring posted data on SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and video analysis, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages based on the prediction results, and means for notifying the generated messages to the user. This enables multifaceted analysis of the user's mental health status from SNS posted data, and enables highly accurate risk prediction and generation of warning and response messages.

[1265] "Means for obtaining data posted on SNS" refers to a function that uses the API of a social networking service (SNS) platform to collect data such as text, images, and videos posted with the user's permission.

[1266] "Means for classifying acquired posted data into text, images, and videos" is a function that detects the format of collected posted data and sorts it into the appropriate module as text data, image data, or video data.

[1267] "Means for analyzing classified data using natural language processing, computer vision, and video analysis" refers to functions for analyzing text data using natural language processing technology (NLP), analyzing image data using computer vision technology, and analyzing video data using video analysis technology.

[1268] "Means for integrating analyzed data and using a generative AI model to predict mental health risks" refers to a function that integrates analyzed text, image, and video data and uses a generative AI model to predict a user's mental health risks.

[1269] The "means for generating a warning message and a response message based on the prediction result" is a function for generating a warning message for the user and a message suggesting specific response methods based on the mental health risk prediction result.

[1270] The "means for notifying the user of the generated message" is a function for sending the generated warning message and corresponding message to the user via a dedicated app, email, or push notification.

[1271] The system of the present invention includes a plurality of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. Specific embodiments for carrying out the present invention are described below.

[1272] Hardware and software used

[1273] Hardware:

[1274] Server: High-performance server (e.g. AWS EC2, Google Cloud Compute Engine)

[1275] Device: Smartphone (e.g. iPhone, Android device)

[1276] software:

[1277] Natural Language Processing algorithms: BERT and GPT-based models

[1278] Computer Vision Technologies: OpenCV, TensorFlow, Keras

[1279] Video analysis algorithms: OpenPose, YOLOv3

[1280] Generative AI models: GPT-3 and the latest generative AI models

[1281] Data collection

[1282] The server collects user post data through the API of the social media platform. Specifically, the server periodically calls the API to retrieve user posts (text, images, videos). This also includes a step to obtain the user's explicit permission. For example, the server uses the Twitter API to call "GET statuses / user_timeline" to collect the latest posts from a specified user.

[1283] Data Classification

[1284] The server classifies the collected data into three modules: text, image, and video. Specifically, it checks the data format and sends the data to the appropriate analysis module. For example, text data is sent to the NLP (natural language processing) module, image data is sent to the computer vision module, and video data is sent to the video analysis module.

[1285] Data analysis

[1286] The server analyzes the data in each module. The specific operations are as follows:

[1287] Text data: The NLP module performs keyword extraction and sentiment analysis using BERT and GPT-based models. For example, it extracts the keyword "tired" and determines the sentiment as "fatigue."

[1288] Image data: The computer vision module uses OpenCV and TensorFlow to perform facial expression recognition, object recognition, and scene analysis, for example, analyzing whether a face in an image has a neutral expression.

[1289] Video data: The video analysis module uses speech recognition and video analysis algorithms (e.g., OpenPose, YOLOv3) to perform multifaceted sentiment analysis, such as analyzing changes in voice tone and facial expressions.

[1290] Mental health risk prediction

[1291] The server integrates the analysis results and uses a generative AI model to predict mental health risks. Specifically, it compares the user's past data with the current analysis results to detect abnormal emotional and behavioral patterns. For example, if the user's posted text frequently contains the word "tired," this will be added up as a risk.

[1292] Generates warnings and response messages

[1293] If the server predicts a high mental health risk, it immediately generates a warning message. Specifically, it generates messages such as "Consult a specialist" or "Take a rest" depending on the risk level. The generative AI model also suggests specific actions that the user can take. For example, it generates a message that reads, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist."

[1294] User Notifications

[1295] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and are selected based on the user's settings. For example, the push notification function of a smartphone app can be used to display a message saying "Consult an expert." However, email notification is also possible.

[1296] Specific examples

[1297] Example 1: For user A

[1298] User A posts on social media, "I've been feeling really tired lately."

[1299] The server collects these posts using the SNS API and sends them to the text analysis module.

[1300] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[1301] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[1302] A generative AI model integrates this data to predict elevated mental health risks.

[1303] The server displays a warning message saying, "Your mental health is at increased risk. Take a rest or consult a professional," along with a suggestion for specific action: "Start practicing mindfulness to reduce stress."

[1304] The device sends this message as a push notification to User A. User A checks the notification and takes appropriate action, such as consulting an expert.

[1305] Prompt Sentence Examples

[1306] "Predict the mental health risk of users who recently posted on social media that they've been feeling very tired lately."

[1307] "Identify the mental health risk from the following post: 'Image: Expressionless face, Text: I've been feeling tired lately' and generate an appropriate warning message."

[1308] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks in users early from SNS posting data and respond promptly.

[1309] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1310] Step 1: Data collection

[1311] The server collects user post data using the API of the SNS platform. Specifically, the server periodically makes API calls to retrieve user posts (text, images, videos) using endpoints such as "GET statuses / user_timeline". The input is the post data obtained as a result of the SNS API call, and the output is the retrieved raw data.

[1312] Step 2: Data Classification

[1313] The server classifies the collected posted data into three types: text, image, and video. The server checks the data format and distributes text data to the NLP module, image data to the computer vision module, and video data to the video analysis module. The input is the collected raw data, and the output is data classified by format.

[1314] Step 3: Text data analysis

[1315] The server processes the text data in a text analysis module. Specifically, it uses BERT or GPT-based natural language processing (NLP) models to extract keywords and perform sentiment analysis. For example, from the text posted by a user saying "I've been feeling tired lately," it extracts the keyword "tired" and recognizes the emotion as "fatigue." The input is text data, and the output is the analyzed text data and extracted sentiment information.

[1316] Step 4: Image data analysis

[1317] The server processes image data using an image analysis module. Specifically, it uses OpenCV and TensorFlow to recognize facial expressions in images and determine their emotions. For example, it performs expressionless face detection and background scene analysis. The input is image data, and the output is analyzed image data and emotional information.

[1318] Step 5: Video data analysis

[1319] The server processes the video data using a video analysis module. Specifically, it uses OpenPose and YOLOv3 to analyze the audio and video in the video and detect emotional changes. For example, it analyzes changes in voice tone and facial expressions. The input is the video data, and the output is the analyzed video data and emotional information.

[1320] Step 6: Data integration and mental health risk prediction

[1321] The server integrates the analysis results obtained from text, images, and videos and uses a generative AI model to predict mental health risks. Specifically, it centralizes this data and compares it with the user's past data to detect abnormal patterns. The input is the analyzed text, image, and video data, and the output is the predicted mental health risk.

[1322] Step 7: Generate warning and response messages

[1323] The server generates a warning message based on the risk prediction result. For example, it generates a message such as "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist," or a message suggesting specific measures to take, such as "Take a break" or "Talk to a friend." The input is the risk prediction result, and the output is the generated warning message and response message.

[1324] Step 8: User Notification

[1325] The device notifies the user of the message received from the server. Notification methods include a dedicated app, email, and push notification, and notifications are sent in the appropriate way depending on the user's settings. For example, the push notification function of a smartphone app can be used to display the generated warning message. The input is the generated warning message and the corresponding message, and the output is the notification sent to the user.

[1326] (Application example 1)

[1327] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1328] In recent years, the spread of social media has led to many people sharing information on a daily basis. However, at the same time, there has been an increase in the number of mental health risks hidden in posts on social media, and there is a need to detect these risks early and take appropriate measures. In addition, there is a problem that it is difficult to respond in real time due to the lack of technology that can instantly notify users of risks using devices such as smartphones or smart glasses.

[1329] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1330] In this invention, the server includes means for acquiring posted data on the SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the generated messages to the user, and means for notifying the user of the warning messages via a smartphone, smart glasses, or similar device. This enables early detection of mental health risks from the content of posts on the SNS and notification to the user in real time.

[1331] "SNS" is an online platform that allows users to share their information and opinions over the Internet.

[1332] "Posted data" refers to digital content such as text, images, and videos posted by users on SNS.

[1333] "Text classification" is the process of analyzing acquired post data based on the content of the strings and sorting them into specific categories.

[1334] "Image classification" is a technology that analyzes images in acquired posting data and classifies them into specific categories based on their content.

[1335] "Video classification" is a technology that analyzes videos in the acquired posting data and classifies them based on their content and characteristics.

[1336] "Natural language processing" is a general term for algorithms and technologies that allow computers to understand and analyze human language.

[1337] "Computer vision" refers to the technology and algorithms used to analyze and understand image and video data.

[1338] "Voice analysis" is a technology that analyzes voice data in digital format and recognizes its content and characteristics.

[1339] "Mental health risks" refer to conditions or symptoms that may have a negative impact on mental health.

[1340] A "warning message" is alert information that notifies the user of a crisis or risk.

[1341] The "response message" is information that suggests specific actions or methods of response to the user who receives the warning message.

[1342] A "generative AI model" is a technology that uses an artificial intelligence model that learns from existing data and experience to predict the characteristics and risks of new data.

[1343] "Notification means" refers to the method or technique for notifying the user of the generated message.

[1344] A "smartphone" is a mobile phone device with advanced computing power and connectivity.

[1345] "Smart glasses" are wearable devices that can integrate and display real and digital information.

[1346] "Device" is a general term for electronic devices that have specific functions or roles.

[1347] The system for implementing the present invention includes a wide range of means for analyzing data posted on SNS and predicting and issuing warnings about mental health risks. A specific method for implementing this system is described below.

[1348] 1. Data Collection

[1349] The server uses the API of the SNS platform to collect user posted data. This requires the user's explicit permission. The collected data is multimodal data such as text, images, and videos.

[1350] 2. Data Classification

[1351] The server categorizes the collected data: text data is sent to a text analysis module, image data to an image analysis module, and video data to a video analysis module.

[1352] 3. Data Analysis

[1353] The server analyzes the data using various analysis modules.

[1354] The text data is analyzed using natural language processing (NLP) algorithms to extract keywords and perform sentiment analysis.

[1355] The image data is subjected to facial expression recognition, object recognition, and scene analysis using computer vision techniques.

[1356] Video data is analyzed using a video analysis algorithm, and multifaceted emotional analysis is performed using voice recognition and video analysis.

[1357] 4. Mental health risk prediction

[1358] The server integrates the analyzed data and predicts mental health risks using a generative AI model that compares it with the user's past data to detect abnormal emotional and behavioral patterns.

[1359] 5. Generating warning and response messages

[1360] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[1361] 6. User Notices

[1362] The device notifies the user of the generated message via a dedicated app, email, or push notification. Additionally, notifications can be sent via smartphones, smart glasses, or similar devices. The user can then review the notifications to understand their mental health status and any necessary actions.

[1363] Specific examples

[1364] Example 1: For user A

[1365] User A posts on social media, "I've been feeling really tired lately." The server collects this post and sends it to a text analysis module. An NLP algorithm extracts the keyword "tired," and sentiment analysis determines the feeling as "fatigue." Computer vision analysis detects that the posted image contains an inconspicuous background and an expressionless face. A generative AI model integrates this data and predicts a high mental health risk. The server generates a warning message for User A saying, "Your mental health risk is increasing. Please take a rest or consult a specialist." The device notifies User A of this message. User A checks the notification and takes appropriate action, such as consulting a specialist.

[1366] Example prompts to input to the generative AI model

[1367] "Analyze the following text and image posted by a user on social media and predict their mental health risk. Text: "I'm so sad", Image: <Image URL>"

[1368] This system makes it possible to detect mental health risks from social media posts early on and provide prompt responses.

[1369] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1370] Step 1:

[1371] Data collection

[1372] The server collects user posted data using the API of the SNS platform. First, with the user's explicit permission, it retrieves text, image, and video data from the SNS using a dedicated app and API key. The input is the user's SNS account information and API key, and the output is the collected SNS posted data.

[1373] Step 2:

[1374] Data Classification

[1375] The server classifies the collected data. The acquired text data is sent to the text analysis module, the image data is sent to the image analysis module, and the video data is sent to the video analysis module. The input is the collected SNS post data, and the output is classified text, image, and video data.

[1376] Step 3:

[1377] Text Data Analysis

[1378] The server analyzes the text data using natural language processing (NLP) algorithms. Specifically, it extracts keywords and performs sentiment analysis. The input is text data, and the output is analyzed emotional information and keywords. For example, the server can detect the emotion "sad" from the posted text "very sad."

[1379] Step 4:

[1380] Image data analysis

[1381] The server analyzes image data using computer vision technology. Specifically, it performs facial expression recognition, object recognition, and scene analysis. The input is image data, and the output is analyzed facial expression information and scene information. For example, neutral faces and dark backgrounds are detected.

[1382] Step 5:

[1383] Video Data Analysis

[1384] The server uses a video analysis algorithm to perform multifaceted emotion analysis using voice recognition and video analysis. The input is video data, and the output is analyzed emotional information and behavioral patterns. For example, if the user's voice tone is low and the video is dark, a report will be generated.

[1385] Step 6:

[1386] Data integration and risk prediction

[1387] The server integrates the analyzed text, image, and video data and uses a generative AI model to predict mental health risks. The input is multiple analyzed data (emotional information, facial expression information, and behavioral patterns), and the output is the predicted mental health risk. For example, abnormal emotional and behavioral patterns are detected by comparing them with past data.

[1388] Step 7:

[1389] Generates warnings and response messages

[1390] The server generates a warning message and a corresponding response message based on the risk prediction result. The input is the mental health risk prediction result, and the output is the generated warning message and corresponding response message. For example, a message such as "Your mental health risk is increasing. Take a rest or consult a specialist" may be generated.

[1391] Step 8:

[1392] User Notifications

[1393] The device notifies the user of the generated message. The notification is via a smartphone, smart glasses, or similar device. The input is the generated message, and the output is a notification displayed on the user's device. The user reviews the notification and receives information about their mental health condition and how to respond.

[1394] This processing step makes it possible to detect mental health risks early from the content of social media posts and notify users in real time.

[1395] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1396] The system of the present invention analyzes data posted on social networking sites to predict and warn of mental health risks, and in particular incorporates an emotion engine to provide a function for recognizing users' emotions in detail. Specific embodiments for implementing the present invention are described below.

[1397] 1. Data Collection

[1398] The server collects user posted data using the API of the SNS platform. After obtaining the user's explicit permission, it acquires multimodal data such as text, images, and videos.

[1399] 2. Data Classification

[1400] The server categorizes the collected data into text, images, and videos, and sends the categorized data to the corresponding analysis module.

[1401] 3. Data Analysis

[1402] The server analyzes the data using various analysis modules.

[1403] The text data is analyzed using natural language processing (NLP) algorithms. Specifically, keyword extraction and sentiment analysis are performed to detect, for example, the keyword "fatigue" and its associated sentiment (fatigue).

[1404] The image data is then subjected to computer vision techniques for facial expression recognition, object recognition, and scene analysis, for example to detect neutral faces and dark backgrounds.

[1405] Video data is analyzed using a video analysis algorithm, which uses voice recognition and video analysis to perform multifaceted emotion analysis, such as detecting low voice tones and little movement.

[1406] 4. Emotion Recognition by Emotion Engine

[1407] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[1408] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions.

[1409] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions, to identify the emotional state in the image.

[1410] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc.

[1411] 5. Mental health risk prediction

[1412] The server integrates the emotional data recognized by the emotion engine and predicts mental health risks using a generative AI model, which compares the data with the user's past data to detect abnormal emotional and behavioral patterns.

[1413] 6. Generating warning and response messages

[1414] If a mental health risk is detected, the server immediately generates a warning message, such as, "Your recent posts have been determined to indicate a high mental health risk. Please consult a specialist." It also generates a response message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[1415] 7. User Notices

[1416] The device will then notify the user of the generated message via a dedicated app, email, or push notification. The user can then review the notification to understand their mental health status and any necessary actions.

[1417] Specific examples

[1418] Example 1: For user A

[1419] User A posts on social media, "I've been feeling really tired lately."

[1420] The server collects the posts and sends them to a text analysis module.

[1421] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[1422] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[1423] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[1424] A generative AI model integrates this data and predicts elevated mental health risks.

[1425] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[1426] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[1427] The above is a specific embodiment of the present invention. This system makes it possible to detect mental health risks from social media posts at an early stage and provide appropriate responses promptly.

[1428] The processing flow will be explained below.

[1429] Step 1:

[1430] A user posts content on an SNS, such as text, images, and videos. For example, a user might post text such as "I've been feeling really tired lately," and attach an image or video.

[1431] Step 2:

[1432] Collecting posted data. The server collects user posted data through the SNS API. This requires explicit permission from the user. The collected data includes text, images, and videos.

[1433] Step 3:

[1434] Data classification: The server classifies the collected posting data into text, images, and videos, and sends the classified data to the corresponding analysis module.

[1435] Step 4:

[1436] Analysis of text data. The server's text analysis module uses natural language processing (NLP) algorithms to analyze the text data. Specifically, it performs keyword extraction and sentiment analysis, detecting, for example, the keyword "fatigue" and its associated sentiment (feeling of fatigue).

[1437] Step 5:

[1438] Image data analysis. The server's image analysis module uses computer vision technology to analyze the image data. Specifically, it performs facial expression recognition, object recognition, and scene analysis to detect, for example, neutral expressions or dark backgrounds.

[1439] Step 6:

[1440] Video data analysis. The server's video analysis module uses a video analysis algorithm to analyze the video data. Specifically, it analyzes emotions through voice recognition and analyzes movements and behaviors within the video. For example, it detects low voice tones and little movement.

[1441] Step 7:

[1442] Emotion recognition by emotion engine. The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[1443] From the text data, the emotion engine analyzes sentences and phrases to recognize positive, negative, and neutral emotions, for example, recognizing negative emotions from the phrase "tired."

[1444] From the image data, the emotion engine analyzes facial expressions and scenes to recognize emotions, such as smiling, sad, or neutral expressions.

[1445] From the video data, the emotion engine analyzes audio and video and recognizes emotions from tone, gestures, movements, etc. For example, a low voice tone and little movement can indicate fatigue.

[1446] Step 8:

[1447] Multimodal data integration: The server integrates the results of text, image, and video analysis to generate a single multimodal dataset that comprehensively represents the user's emotional state.

[1448] Step 9:

[1449] Mental health risk prediction. The server's generated AI model uses multimodal data to predict the user's mental health risk. For example, if a pattern of "fatigue" and "expressionlessness" persists compared to past data, it will determine that the risk is high.

[1450] Step 10:

[1451] Generate a warning message. If the server determines that the risk level is high, it generates a warning message. For example, it creates a message that reads, "Your recent posts have been determined to pose a high risk to your mental health. Please consult a professional."

[1452] Step 11:

[1453] Generating a response message: The server generates a message suggesting specific actions to take, such as "take a rest," "talk to a friend," or "consult a doctor."

[1454] Step 12:

[1455] Message notification: The server sends the generated warning message and corresponding message to the device. The message is notified to the user via a dedicated app, email, or push notification.

[1456] Step 13:

[1457] Message review and action: Users can review notifications, understand their mental health status, take suggested actions, and seek professional help if necessary.

[1458] Example 2

[1459] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1460] Conventional systems have had difficulty accurately predicting users' mental health risks from data posted on social media and generating appropriate warning and response messages. Furthermore, it is not easy to comprehensively analyze a variety of data (text, images, and videos) with explicit permission from the user. This makes it difficult to address users' mental health issues early, and there is a need for improved accuracy in risk prediction.

[1461] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1462] In this invention, the server includes means for acquiring posted data on SNS, means for classifying the posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and voice analysis, means for performing detailed emotion recognition on the analyzed text, image, and video data using an emotion engine, means for integrating the analyzed data and predicting mental health risks using a generative AI model, means for generating warning messages and response messages, and means for notifying the generated messages to the user. This makes it possible to accurately predict a user's mental health risks from the posted data on SNS and provide early and appropriate responses.

[1463] "SNS posting data" refers to content such as text, images, and videos posted by users on social networking services (SNS).

[1464] "Text" refers to data containing text information posted by users on SNS.

[1465] "Images" are data containing visual content posted by users on social networking sites, including photographs and illustrations.

[1466] "Video" refers to data containing content accompanied by motion and sound that users post on social networking sites.

[1467] "Natural language processing" is a technology that uses computers to analyze and understand human language, and includes text data analysis, keyword extraction, sentiment analysis, and more.

[1468] "Computer vision" is a technology that allows computers to analyze and understand visual information such as images and videos.

[1469] "Voice analysis" is a technology that analyzes voice data and performs functions such as converting it into text, analyzing emotions, and recognizing speech.

[1470] An "emotion engine" is a technology that comprehensively analyzes data such as text, images, and videos to recognize the user's emotional state in detail.

[1471] A "generative AI model" is an artificial intelligence model that uses machine learning technology to recognize specific patterns from input data and make predictions or generate results.

[1472] "Mental health risk" refers to the possibility that a user may have problems with their mental health, and is predicted based on abnormal emotions and behavioral patterns from analytical data.

[1473] A "warning message" is a message that notifies the user that their mental health risk is increasing.

[1474] The "response message" is a message that suggests specific ways to respond to a user whose mental health risk is increasing.

[1475] "Means of notification" refers to technology for transmitting the generated warning message and response message to the user, and includes dedicated apps, email, push notifications, etc.

[1476] The system of the present invention analyzes data posted on social media platforms to predict and warn of mental health risks. Specifically, the server uses the social media platform's API to collect user posted data (text, images, and videos), and classifies and analyzes this data to predict mental health risks. Furthermore, an emotion engine is incorporated to recognize users' emotions in detail, and appropriate warning and response messages are generated and sent.

[1477] Data collection

[1478] The server collects user posting data using the APIs of social networking platforms such as Twitter and Facebook. Multimodal data such as text, images, and videos are collected with the user's explicit permission. For example, if a user posts "I've been feeling tired lately," the text data and attached image and video data are also collected.

[1479] Data Classification

[1480] The server categorizes the collected posting data into three types: text, images, and videos. For example, a photo of a non-smiling face posted along with a text message saying "I'm not feeling well" or a video in a dark room would be classified separately. This ensures that data is properly sorted according to each analysis module.

[1481] Data analysis

[1482] The server analyzes the data using the following analysis modules:

[1483] The text data is then subjected to keyword extraction and sentiment analysis using NLP (Natural Language Processing) algorithms, for example, to detect the keyword "fatigue" and the associated "feeling of fatigue."

[1484] The image data is then subjected to facial expression recognition and scene analysis using computer vision techniques, such as detecting neutral facial expressions and dark backgrounds as features.

[1485] Video data is analyzed using voice recognition and video analysis to detect voice tone and lack of movement, for example, low voice tone and lack of movement.

[1486] Emotion recognition by emotion engine

[1487] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The emotion engine works as follows:

[1488] From the text data, phrase analysis is performed to recognize positive, negative, and neutral emotions. For example, a post saying "I've been feeling tired lately" can be interpreted as a negative emotion.

[1489] Emotions are recognized from image data by analyzing facial expressions and scenes, for example, detecting expressionless faces and dark scenes.

[1490] Video data is analyzed for audio and video to detect low tones and little movement.

[1491] Mental health risk prediction

[1492] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risk. This model compares a user's past emotional data with their current data to detect abnormal emotional and behavioral patterns. For example, if a user's recent posts show an abnormally negative trend compared to past data, it predicts a high mental health risk.

[1493] Generates warnings and response messages

[1494] If the server predicts a high mental health risk, it generates a warning message and a corresponding message, such as "Your mental health is at increased risk. Take a rest or consult a professional."

[1495] User Notifications

[1496] The device will notify the user of the warning message and corresponding message sent from the server via a dedicated app, email, or push notification. Users can check the notification, understand their mental health risk, and take necessary action.

[1497] Specific examples

[1498] Example 1: For user A

[1499] User A posts on social media, "I've been feeling really tired lately."

[1500] The server collects the posts and sends them to a text analysis module.

[1501] The NLP algorithm extracts the keyword "tired" and determines the feeling as "fatigue" through sentiment analysis.

[1502] Computer vision analysis detects whether posted images contain unobtrusive backgrounds or neutral faces.

[1503] The emotion engine recognizes emotions such as "fatigue" and "expressionless" in detail from text and image data.

[1504] A generative AI model integrates this data and predicts elevated mental health risks.

[1505] The server generates a warning message to User A saying, "Your mental health is at increased risk. Take a rest or consult a professional."

[1506] The terminal notifies this message to User A. User A checks the notification and takes action such as consulting an expert.

[1507] Example prompts for generative AI models

[1508] "Please tell me the system procedure for analyzing a user's emotional state from social media posts and predicting mental health risks."

[1509] "Please explain each processing step of the system that analyzes emotions based on user posts on social media and predicts mental health risks."

[1510] The above is a specific embodiment for carrying out the present invention. This system makes it possible to detect mental health risks of users early from posts on SNS and quickly provide appropriate responses.

[1511] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1512] Step 1: Data collection

[1513] The server uses the API of the SNS platform to collect user posted data. API authentication information (API key and secret) and user ID are required as input. The acquired data includes text, images, and videos. Specifically, the server sends an API request to the SNS platform and stores the user posted data in a temporary database.

[1514] Step 2: Data Classification

[1515] The server classifies the collected posted data into text, images, and videos. The input requires the multimodal data collected in step 1. The output is data classified into text data, image data, and video data. Specifically, the server checks the data format and distributes the data to the appropriate analysis module.

[1516] Step 3: Data analysis

[1517] The server analyzes data using various analysis modules. Text data, image data, and video data are required as input. Analysis results (keywords, emotional state, facial expression, voice tone, etc.) are obtained as output. The specific operation is as follows:

[1518] NLP algorithms are used to extract keywords and analyze sentiment from text data. For example, the keyword "tired" is extracted and the sentiment "fatigue" is determined.

[1519] Facial expression recognition is performed on image data using computer vision techniques. For example, neutral faces are detected.

[1520] Speech recognition and video analysis are performed on video data to detect voice tones and lack of movement. Example: Low voice tones and lack of movement are detected.

[1521] Step 4: Emotion Recognition with the Emotion Engine

[1522] The server sends the analyzed text, image, and video data to the emotion engine to recognize the user's emotions in detail. The analyzed data from step 3 is required as input. The output is a detailed emotional state (positive, negative, neutral). Specifically, the emotion engine performs phrase analysis, facial expression analysis, and voice analysis to recognize emotions from each media.

[1523] Step 5: Mental health risk prediction

[1524] The server integrates the emotional data recognized by the emotion engine and uses a generative AI model to predict mental health risks. Detailed emotional data and past emotional data are required as input. The output is a predicted mental health risk. Specifically, the system inputs emotional data into the generative AI model and performs risk assessment. In particular, it compares past data with current data to detect abnormal emotional and behavioral patterns.

[1525] Step 6: Generate warning and response messages

[1526] The server generates a warning message and a response message if a high mental health risk is predicted. The required input is the risk assessment result. The output is a warning message and a response message. Specifically, based on the risk assessment, the server generates a warning message such as "Take a rest or consult a specialist" and a message suggesting specific response methods such as "Talk to a friend" or "Consult a doctor."

[1527] Step 7: User Notification

[1528] The device notifies the user of the warning message and response message sent from the server. The generated message is required as input. The output is the completion of the notification to the user. The specific operation is to display the message via a dedicated app, email, or push notification. The user can then check the notification and take appropriate action.

[1529] (Application example 2)

[1530] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1531] In modern society, with the spread of self-driving vehicles, mental health issues among drivers are on the rise, and there is a need for a system that can monitor their mental state in real time and encourage appropriate responses. However, conventional technologies predict mental health risks solely from data posted on social media, making it difficult to reflect real-time conditions while driving. In addition, warning messages are not generated and notified quickly, making it difficult for drivers to recognize problems in a timely manner and take appropriate measures. These issues need to be resolved.

[1532] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring posted data on an SNS, means for classifying the acquired posted data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and speech analysis, means for predicting mental health risks from the analyzed data, means for generating warning messages and response messages based on the prediction results, means for notifying the user of the generated messages, means for analyzing data collected from sensors in the autonomous vehicle and evaluating the driver's mental state, and means for displaying a warning message using the vehicle's dashboard display system based on the evaluation results. This enables real-time monitoring of the driver's mental state while driving, and prompt warnings and response suggestions.

[1533] "SNS" is an abbreviation for social networking service, a platform for users to communicate and share information online.

[1534] "Posted data" refers to all information, including text, images, and videos, posted by users on SNS.

[1535] A "server" is a computer system that processes and manages information on a network.

[1536] "Natural language processing" refers to techniques and approaches that enable computers to understand and generate human language.

[1537] "Computer vision" is a technology that allows computers to analyze digital images and videos and understand visual information.

[1538] "Voice analysis" is a technology that analyzes voice data and extracts its content and characteristics.

[1539] "Mental health risk" refers to the possibility that an individual's mental health will deteriorate.

[1540] A "warning message" is a message that the system uses to notify the user of risks or situations that require caution.

[1541] A "response message" is a notification message that suggests specific actions or countermeasures for responding to a warning.

[1542] An "autonomous vehicle" is a vehicle that can be driven and navigated automatically without the need for human operation.

[1543] A "sensor" is a device that detects changes in the physical environment and acquires that information as electronic data.

[1544] A "dashboard" is the part of a vehicle that contains a display and interface through which the driver can view information.

[1545] This invention is a system that analyzes data posted on social media to predict and warn drivers about mental health risks, and is particularly applicable to self-driving vehicles. The system is configured as follows:

[1546] Hardware

[1547] Camera: Installed inside the vehicle to capture the driver's facial expressions and movements.

[1548] Microphone: Installed inside the vehicle to collect the driver's voice and audio.

[1549] Smart glasses: Worn by the driver, they track eye movements and gaze in real time.

[1550] Dashboard display: A display for showing warning and response messages.

[1551] software

[1552] SNS API: Retrieves post data from social media platforms.

[1553] Natural language processing (NLP) engine: Analyzes text data, extracts keywords, and performs sentiment analysis.

[1554] Computer vision tools: Analyze image data and perform facial expression and object recognition.

[1555] Video analysis algorithm: Analyzes video data and detects emotions from audio and video.

[1556] Emotion engine: Integrates various analytical data to recognize the driver's emotional state in detail.

[1557] Generative AI model: Predicting mental health risks based on integrated emotional data.

[1558] Notification system: Notifies the driver of warning and response messages via the dashboard display.

[1559] Program processing

[1560] The server collects driver posting data using SNS APIs. The collected data is categorized into text, images, and videos, and each data is sent to the corresponding analysis module. The NLP engine analyzes the text data to extract keywords and perform sentiment analysis. The computer vision tool analyzes the image data to recognize facial expressions and scenes. The video analysis algorithm analyzes the video data and extracts emotions from audio and video.

[1561] The analyzed data is integrated into an emotion engine to recognize the driver's emotional state in detail. The generative AI model uses this emotional data to predict mental health risks and immediately generates a warning message if the risk is high. For example, it might say, "Your mental health risk is increasing. Please take a break or consult a specialist." It also generates a response message suggesting specific ways to respond.

[1562] These messages are displayed on the dashboard display, allowing the driver to check their own mental health status in real time and take necessary action. This system makes it possible to detect mental health risks while driving early and provide appropriate responses quickly.

[1563] Specific examples

[1564] For example, if Driver B posts on social media while driving, "I'm tired right now, but I'm driving hard," the server collects this post, extracts the keyword "tired" using an NLP engine, and determines the state as "fatigue" using sentiment analysis. Furthermore, the vehicle's camera detects expressionless faces, and the microphone recognizes monotone voices. This data is integrated, and the sentiment engine precisely recognizes the emotions of "fatigue" and "expressionless," and the generative AI model predicts a high mental health risk. The system generates a warning message saying, "Your mental health risk is increasing. You should take a break," and displays this on the dashboard display. Driver B sees this message and takes action to rest in a safe place.

[1565] Prompt Sentence Examples

[1566] Input prompt: "Based on recent posts, predict risks and generate a warning message."

[1567] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1568] Step 1:

[1569] The server collects driver posting data using SNS API. The collected data is in the form of text, images, and videos, and is executed with the user's explicit permission. The SNS posting data is used as input, and it is output in the form of classified data.

[1570] Step 2:

[1571] The server classifies the collected data into text, images, and videos. The input here is the various posted data collected in step 1, and the classified data format (text, images, videos) is output. This classification is done automatically.

[1572] Step 3:

[1573] The server analyzes the classified data. Specifically, text data is analyzed using a natural language processing (NLP) engine to extract keywords and perform sentiment analysis. Image data is analyzed using computer vision tools to recognize facial expressions and scenes. Video data is analyzed using video analysis algorithms to extract emotions from audio and video. The input is the data classified in step 2, and the analyzed emotional data is the output.

[1574] Step 4:

[1575] The server integrates the analyzed data and performs detailed emotion recognition using an emotion engine. Here, the analysis results of each piece of text, image, and video data are input, and the integrated emotion data is output. The emotion engine recognizes the emotional state of each piece of data in detail and provides integrated emotion information.

[1576] Step 5:

[1577] The server uses a generative AI model based on the integrated emotion data to predict mental health risks. The input is the integrated emotion data, and the output is the predicted mental health risk. This prediction is made by comparing it with past data to detect abnormal emotion and behavior patterns.

[1578] Step 6:

[1579] If the server predicts a high mental health risk, it immediately generates a warning message and a corresponding message. The input here is the predicted mental health risk, and the message output is something like "Your mental health risk is increasing. Take a rest or consult a specialist."

[1580] Step 7:

[1581] The terminal notifies the driver of the generated message. Specifically, it displays warning messages and response messages in real time on the dashboard display so that the driver can check them. The input is the message generated in step 6, and the output is the displayed message. The driver can check this message and take safety measures.

[1582] For example, if Driver B posts on social media, "I'm tired right now, but I'm driving hard," the camera will detect an expressionless face and the microphone will detect a monotone voice. These data will be integrated to recognize the emotions of "fatigue" and "expressionless," and the generative AI model will predict an increase in mental health risk and display an appropriate warning message.

[1583] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1584] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1585] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1586] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1587] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1588] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1589] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1590] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1591] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1592] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1593] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1594] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1595] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1596] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1597] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1598] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1599] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1600] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1601] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1602] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1603] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1604] The following is further disclosed regarding the above embodiment.

[1605] (Claim 1)

[1606] A means of acquiring data posted on SNS,

[1607] A means to classify the acquired post data into text, images, and videos,

[1608] means for analyzing the classified data using natural language processing, computer vision, and speech analysis;

[1609] A means of predicting mental health risks from the analyzed data;

[1610] means for generating a warning message and a corresponding message based on the prediction result;

[1611] means for notifying a user of the generated message;

[1612] A system including:

[1613] (Claim 2)

[1614] 2. The system according to claim 1, further comprising means for obtaining explicit permission from a user when acquiring data posted on an SNS.

[1615] (Claim 3)

[1616] 10. The system of claim 1, further comprising means for integrating the analyzed data and predicting mental health risk using a generative AI model.

[1617] "Example 1"

[1618] (Claim 1)

[1619] A means of acquiring data posted on SNS,

[1620] A means to classify the acquired post data into text, images, and videos,

[1621] a means for analyzing the classified data using natural language processing, computer vision, and video analysis;

[1622] A means of integrating the analyzed data and using a generative AI model to predict mental health risks; and

[1623] means for generating a warning message and a corresponding message based on the prediction result;

[1624] means for notifying a user of the generated message;

[1625] A system including:

[1626] (Claim 2)

[1627] 2. The system according to claim 1, further comprising means for obtaining explicit permission from a user when acquiring data posted on an SNS.

[1628] (Claim 3)

[1629] The system of claim 1, further comprising means for notifying a user of the generated warning message and the corresponding message via a dedicated app, email, or push notification.

[1630] "Application Example 1"

[1631] (Claim 1)

[1632] A means of acquiring data posted on SNS,

[1633] A means to classify the acquired post data into text, images, and videos,

[1634] means for analyzing the classified data using natural language processing, computer vision, and speech analysis;

[1635] A means of predicting mental health risks from the analyzed data;

[1636] means for generating a warning message and a corresponding message based on the prediction result;

[1637] means for notifying a user of the generated message;

[1638] a means for communicating the warning message to a smartphone, smart glasses, or similar device;

[1639] A system including:

[1640] (Claim 2)

[1641] 2. The system according to claim 1, further comprising means for obtaining explicit permission from a user when acquiring data posted on an SNS.

[1642] (Claim 3)

[1643] 10. The system of claim 1, further comprising means for integrating the analyzed data and predicting mental health risk using a generative AI model.

[1644] "Example 2: Combining Emotion Engines"

[1645] (Claim 1)

[1646] A means of acquiring data posted on SNS,

[1647] A means to classify the acquired post data into text, images, and videos,

[1648] means for analyzing the classified data using natural language processing, computer vision, and speech analysis;

[1649] A means to perform detailed emotion recognition using an emotion engine on the analyzed text, image, and video data;

[1650] A means of integrating the analyzed data and using a generative AI model to predict mental health risks; and

[1651] means for generating a warning message and a corresponding message;

[1652] means for notifying a user of the generated message;

[1653] A system including:

[1654] (Claim 2)

[1655] 2. The system according to claim 1, further comprising means for obtaining explicit permission from a user when acquiring data posted on an SNS.

[1656] (Claim 3)

[1657] 10. The system of claim 1, further comprising means for integrating the analyzed data and predicting mental health risk using a generative AI model.

[1658] "Application example 2 when combining emotion engines"

[1659] (Claim 1)

[1660] A means of acquiring data posted on SNS,

[1661] A means to classify the acquired post data into text, images, and videos,

[1662] means for analyzing the classified data using natural language processing, computer vision, and speech analysis;

[1663] A means of predicting mental health risks from the analyzed data;

[1664] means for generating a warning message and a corresponding message based on the prediction result;

[1665] means for notifying a user of the generated message;

[1666] A means of analyzing data collected from sensors in autonomous vehicles to assess the mental state of the driver; and

[1667] means for displaying a warning message using a dashboard display system of the vehicle based on the evaluation result;

[1668] A system including:

[1669] (Claim 2)

[1670] 2. The system according to claim 1, further comprising means for obtaining explicit permission from a user when acquiring data posted on an SNS.

[1671] (Claim 3)

[1672] 10. The system of claim 1, further comprising means for integrating the analyzed data and predicting mental health risk using a generative AI model. [Explanation of symbols]

[1673] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of acquiring data posted on SNS, A means to classify the acquired post data into text, images, and videos, means for analyzing the classified data using natural language processing, computer vision, and speech analysis; A means of predicting mental health risks from the analyzed data; means for generating a warning message and a corresponding message based on the prediction result; means for notifying a user of the generated message; A system including:

2. The system according to claim 1 , further comprising means for obtaining explicit permission from a user when acquiring data posted on an SNS.

3. The system of claim 1, further comprising means for integrating the analyzed data and predicting mental health risk using a generative AI model.

Images