system

Abstract

Provide a system. 【Solution means】 An acquisition means for acquiring acoustic information and visual information, A transmission means for transmitting the acquired acoustic information and visual information via a communication protocol, An evaluation means for analyzing the received acoustic information with a natural language processing engine and evaluating the risk of the speech content, An analysis means for analyzing the emotional state from the visual information using a facial expression recognition algorithm, A generation means for determining the degree of risk based on a harassment index and generating feedback information, A notification means for notifying the user of the generated feedback information, A storage means for storing the analysis information and the history information, A training means for retraining a machine learning model based on the stored information, A means having a function of monitoring conversations within a home, detecting specific phrases to assist smooth communication, and providing feedback, A system including the above.

Description

Technical Field

[0005]

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, and includes steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a character of the chatbot, 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

Summary of the Invention

Problems to be Solved by the Invention

[0004] In modern society, there are various types of harassment, and particularly, problems have arisen where communication in the workplace or public places has become difficult or stagnant. Since the perception of such harassment is highly involved with the subjective feelings of the victim, it is difficult to objectively and in real-time determine and prevent harassment. Also, due to the lack of clear criteria, there is a risk of deterioration in the quality of communication. To solve this problem, a system that can objectively determine the risk of harassment and immediately provide feedback to the user is required.

Means for Solving the Problems

[0005] This invention provides a system that acquires audio and video data in real time, transmits it securely, and analyzes the received data using natural language processing and facial expression recognition algorithms. This system quantifies harassment risk and evaluates the risk level, enabling it to generate and notify users of appropriate feedback messages. Furthermore, it continuously improves the system's accuracy by saving analysis data and history, and retraining the machine learning model based on this data. In addition, privacy protection measures, such as encryption during data transmission, are implemented to provide a secure environment for use. This facilitates smoother communication and prevents harassment.

[0006] "Acquisition means" refers to the functions of devices and software for collecting audio and video data in real time.

[0007] "Transmission means" refers to the function of sending collected audio and video data to a server using a secure communication protocol.

[0008] The "evaluation method" is a function that uses a natural language processing engine to analyze the risks of the content of a statement based on the received data, and then evaluates the results.

[0009] The "analysis means" refers to a processing function that uses an algorithm to recognize the user's emotional state and facial expressions from video data received along with audio data.

[0010] The "generation method" is a function that determines the risk level based on the evaluated harassment score and generates a feedback message to notify the user.

[0011] A "notification method" is a function that informs the user of the generated feedback message visually or audibly.

[0012] "Storage method" refers to a function that records the analyzed data and user feedback history for use in subsequent analysis and model improvement.

[0013] A "training method" is a function that uses stored data to retrain a machine learning model and improve the accuracy and performance of the system. [Brief explanation of the drawing]

[0014] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14]It is a sequence diagram showing the processing flow of a data processing system in Application Example 2 when a sentiment engine is combined.

Embodiments for Carrying Out the Invention

[0015] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0016] First, the terms used in the following description will be explained.

[0017] In the following embodiments, a processor with a reference numeral (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of a plurality of arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of a plurality of types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0018] In the following embodiments, a RAM (Random Access Memory) with a reference numeral is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0019] In the following embodiments, a storage with a reference numeral is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, and the like.

[0020] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0021] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0022] [First Embodiment]

[0023] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0024] As shown in Figure 1, the 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.

[0025] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0026] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0027] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and 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.

[0028] 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0029] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0030] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0032] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0033] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0034] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0035] The system of the present invention acquires audio and video data in real time, analyzes it to assess the risk of harassment, and provides feedback to the user.

[0036] In this embodiment, the terminal first uses a microphone and camera to acquire the user's voice and video as data. This data is transmitted to the server in an encrypted format. To protect privacy, advanced encryption protocols are used for communication during data transmission.

[0037] The server analyzes the received data. First, it converts audio data into text using natural language processing (NLP) technology, and then measures the likelihood of harassment based on the text structure and content. Simultaneously, it analyzes the user's emotional state using facial recognition algorithms based on video data. This analysis allows for a multifaceted evaluation of the user's intentions and how they were perceived.

[0038] Next, the server scores the harassment risk based on the analysis results and determines the risk level (low, medium, or high) based on that score. It then generates an appropriate feedback message accordingly. This generated feedback is returned to the terminal in real time and notified to the user. Based on the feedback, the user can immediately review their own words and actions.

[0039] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and facial expressions and sends them to the server. The server assigns a high risk score to this comment, citing gender bias, and notifies the user through the device that "this comment may be inappropriate."

[0040] Furthermore, the server stores a history of all analysis and feedback, and uses this to retrain the machine learning model, improving accuracy and reliability. This process allows the system to continuously provide intuitive and useful feedback to users, thereby preventing harassment and improving the quality of communication.

[0041] The following describes the processing flow.

[0042] Step 1:

[0043] The device uses a microphone and camera to acquire the user's voice and video data. The data is captured in real time and temporarily stored by the system.

[0044] Step 2:

[0045] The device compresses the acquired audio and video data and sends it to the server in a highly encrypted format. This enhances data security.

[0046] Step 3:

[0047] The server uses a natural language processing (NLP) engine to convert the received audio data into text. Next, it identifies specific keywords and language patterns from the analyzed text to assess the potential for harassment.

[0048] Step 4:

[0049] The server analyzes the received video data using a facial recognition algorithm to evaluate the user's emotional state. This information is combined with text analysis results to perform a comprehensive risk assessment.

[0050] Step 5:

[0051] The server analyzes audio and video data to score the harassment risk and classifies it into three risk levels (low, medium, high). It then generates a feedback message corresponding to this risk level.

[0052] Step 6:

[0053] The server sends the generated feedback message to the terminal.

[0054] Step 7:

[0055] The device notifies the user of the generated feedback message in real time. This gives the user an opportunity to review their own words and actions.

[0056] Step 8:

[0057] The server stores the analysis results and user feedback history in a database.

[0058] Step 9:

[0059] The server uses the stored data to retrain the machine learning model and improve the system's accuracy. This process is performed periodically.

[0060] (Example 1)

[0061] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0062] In today's communication environment, it is essential to recognize the risks of harassment early and for users to appropriately avoid them. However, conventional technologies do not adequately provide efficient methods for detecting harassment using audio and video, and there is a lack of mechanisms to provide real-time feedback on inappropriate remarks and actions that users may unconsciously make. As a result, the detection and correction of harassment is delayed, hindering the improvement of the quality of communication.

[0063] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0064] In this invention, the server includes acquisition means for acquiring acoustic and image data, transmission means for transmitting this data via an information transmission protocol, and analysis means that use a natural language processing engine and image recognition algorithms to analyze the received data, assess the risks, and provide feedback. This enables early detection of harassment risks, and allows users to receive real-time feedback, enabling them to consciously and quickly review their behavior.

[0065] "Audio data" refers to digital or analog data containing audio information, acquired for the purpose of transmitting information based on sound waves.

[0066] "Image data" refers to digital or analog data containing visual information acquired by cameras or other imaging devices.

[0067] An "information transmission protocol" is a set of rules and procedures for communicating data securely and efficiently.

[0068] A "natural language processing engine" is a set of algorithms that analyze speech or text data to support the understanding and generation of natural language used by humans.

[0069] An "image recognition algorithm" is an algorithm that analyzes specific patterns or features within an image to extract and recognize information.

[0070] A "harassment evaluation index" is a standard or criterion used to quantify the possibility and degree of harassment.

[0071] A "computer learning model" is a framework of algorithms that learns patterns from data and performs predictions and classifications.

[0072] In an embodiment of the present invention, a terminal first acquires acoustic and image data using a microphone and a camera. This includes commonly used electronic devices, voice recognition microphones, and HD cameras. This information is encrypted in real time and transmitted to a server via an information transmission protocol.

[0073] Next, the server converts the received audio data into text data using a natural language processing engine. This engine includes, for example, speech recognition software. The server then analyzes the text data and assesses the potential risks based on the content of the speech. Simultaneously, the server analyzes the image data using an image recognition algorithm to determine the user's emotional state. This image recognition algorithm utilizes facial recognition software.

[0074] Based on these analysis results, the server generates harassment evaluation metrics and calculates a risk score in real time. If a high risk is determined, the AI ​​model generates an appropriate feedback message for the user and notifies them through their device.

[0075] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and video information and sends it to the server. The server determines that the comment contains gender bias and scores it as high risk. It then sends feedback to the device stating, "This comment may be inappropriate," and notifies the user. This notification allows the user to immediately review their comment.

[0076] An example of a prompt message would be: "Voice: This job is probably impossible for a woman. Emotional state: Neutral. Please analyze."

[0077] This system allows users to recognize the risk of harassment they may have unintentionally caused early on and have the opportunity to improve it. Furthermore, the server stores all analysis results and feedback history, and uses this data to retrain the computer learning model, continuously improving the overall accuracy and reliability of the system.

[0078] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0079] Step 1:

[0080] The device acquires audio and image data from the user. It records audio with a microphone and captures images with a camera. This acquired data serves as input. The data is encrypted and sent to the server using an information transmission protocol. The output consists of encrypted audio and image data.

[0081] Step 2:

[0082] The server inputs the received audio data into a natural language processing engine, which converts it into text data. Here, speech recognition technology is used to analyze the language from the audio waveform and transcribe it into text. This process outputs the audio content in text format.

[0083] Step 3:

[0084] Simultaneously, the server uses an image recognition algorithm to input image data and analyze the user's emotional state. During the analysis process, it captures feature points of facial expressions, identifies subtle changes in expression, and outputs the emotional state (e.g., joy, anger, sadness).

[0085] Step 4:

[0086] The server integrates text data and image sentiment analysis results, and uses this information to calculate a harassment assessment index. Specifically, it utilizes a generative AI model to input text and sentiment information and score the risk level of the remarks. The output is a numerical score of harassment risk.

[0087] Step 5:

[0088] The server generates feedback information based on the calculated risk score. Using a generation AI model, it creates text messages corresponding to the risk level, and the content of those messages is output.

[0089] Step 6:

[0090] The server sends the generated feedback information to the terminal. The terminal notifies the user of the received feedback, allowing the user to modify their actions based on it. Real-time notifications to the user are output.

[0091] Step 7:

[0092] The server stores all analysis results and feedback history. This dataset is used to retrain the computer learning model, continuously improving the system's accuracy and reliability. Data storage and model training are the outputs.

[0093] (Application Example 1)

[0094] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0095] In family communication, unintentional harassment and inappropriate remarks can damage relationships with family members and guests. Preventing such problems and maintaining a smooth and harmonious communication environment is crucial. However, currently, there is a lack of means to detect these problems in real time and provide feedback. Therefore, there is a need for a system that can address this challenge.

[0096] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0097] In this invention, the server includes acquisition means for acquiring acoustic and visual information; transmission means for transmitting the acquired acoustic and visual information via a communication protocol; evaluation means for analyzing the received acoustic information with a natural language processing engine and evaluating the risk of the content of the speech; analysis means for analyzing emotional states from visual information using an facial expression recognition algorithm; generation means for determining the degree of risk based on harassment indicators and generating feedback information; notification means for notifying the user of the generated feedback information; storage means for storing analysis information and history information; training means for retraining a machine learning model based on the stored information; and means having a function to monitor conversations within the home, detect specific phrases to support smooth communication, and provide feedback. This makes it possible to improve the quality of communication within the home.

[0098] "Acoustic information" refers to data used to acquire ambient sounds and speech.

[0099] "Visual information" refers to video data acquired through a camera.

[0100] A "communication protocol" is a set of rules that define how data can be transmitted securely.

[0101] A "natural language processing engine" is a technology that converts audio data into text and analyzes its content.

[0102] A "facial expression recognition algorithm" is a method for analyzing a user's emotional state from video data.

[0103] A "harassment index" is a set of values ​​used as a standard for evaluating the risks of words and actions.

[0104] "Feedback information" refers to advice and warning messages provided to users based on the analysis results.

[0105] A "machine learning model" is a collection of algorithms used to perform analysis and predictions based on data.

[0106] "Monitoring conversations within the household" means monitoring the speech and actions of members of the household to detect potentially dangerous behaviors.

[0107] This invention is a system designed to improve the quality of communication within the home. To acquire acoustic and visual information, the server uses a terminal equipped with a high-performance microphone and camera. This allows for the acquisition of acoustic and video data in real time. This data is transmitted to the server via a communication protocol. The communication protocol uses encryption technologies such as TLS / SSL to ensure secure data transmission.

[0108] On the server, received audio data is converted into text using a natural language processing engine (e.g., the spaCy library) and the potential danger of the spoken content is evaluated. For visual information, the user's emotional state is analyzed using an facial recognition algorithm based on OpenCV. Based on these analysis results, a harassment index is calculated and the degree of danger is evaluated.

[0109] Based on the assessed risk level, the server generates feedback information. This generated feedback information is then notified to the user. Since the feedback information is presented in real time through the device's display or speaker, the user can immediately review their own statements and actions.

[0110] Furthermore, the server stores analysis information and feedback history. Based on this stored information, the machine learning model can be retrained to improve the accuracy and reliability of the feedback. By detecting specific phrases in conversations within the home and providing appropriate feedback, it supports smoother communication.

[0111] For example, if someone says "My opinion might be considered boring" during a family dinner, the server can pick up on this phrase, analyze the emotion behind it, and then provide feedback such as, "Why not try trusting yourself more and sharing your opinion?"

[0112] An example of a prompt to input into a generative AI model is, "Please describe how to assess the harassment risk in conversations within the family and provide feedback on specific areas for improvement."

[0113] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0114] Step 1:

[0115] The device uses a high-performance microphone and camera to acquire acoustic and visual information within the home. This input consists of audio and video data collected in real time. The device stores this data in a buffer and prepares it for subsequent processing.

[0116] Step 2:

[0117] The terminal transmits the acquired audio and visual information to the server via a secure communication protocol using encryption technologies such as TLS / SSL. The input data is raw audio and video, and the output is this data encrypted and sent to the server. The server decrypts the received data and makes it analyzable.

[0118] Step 3:

[0119] The server converts the received audio data into text using a natural language processing engine (e.g., spaCy). The input for this step is audio data, and the output is the generated text data. The server then performs speech recognition and processes the conversation content as textual information.

[0120] Step 4:

[0121] The server processes the received visual information using an OpenCV-based facial recognition algorithm to analyze the user's emotional state. The input for this step is video data, and the output is numerical data representing emotions. The server analyzes the user's facial expressions and detects emotional patterns.

[0122] Step 5:

[0123] The server integrates natural language processing results and facial recognition results, and uses this to evaluate the potential danger of the spoken content. The input is text data and emotion data, and the output is a danger score as a harassment index. The server analyzes the fused data to score the potential harassment risk.

[0124] Step 6:

[0125] The server generates feedback information based on the risk score and sends it to the terminal. The input is the risk score, and the output is a feedback message to notify the user. The server then forms an appropriate advice or warning message.

[0126] Step 7:

[0127] The terminal notifies the user of feedback information received from the server via its display or speaker. The input is the feedback message, and the output is the transmission of information visually or audibly. This allows the user to immediately review their own actions.

[0128] Step 8:

[0129] The server stores all analysis results and feedback history in a database. The inputs are the analysis data and feedback history, and the output is the recorded data. The server stores this data for future retraining.

[0130] Step 9:

[0131] The server retrains a machine learning model based on stored data. The input is the stored historical data, and the output is the improved model. The server periodically updates the model to improve analysis accuracy and increase the reliability of the feedback.

[0132] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0133] This invention is a system that acquires user voice and video data, assesses the risk of harassment based on this data, and provides immediate feedback to the user. The system further incorporates an emotion engine to recognize the user's emotions.

[0134] First, the device uses its microphone and camera to capture data on the user's conversation and facial expressions. This data is securely processed using encryption technology and sent to the server. The transmitted data is guaranteed to be encrypted to ensure the user's privacy.

[0135] Next, the server converts the received audio data into text using natural language processing (NLP) technology. Based on this text information, it analyzes specific keywords and language patterns to assess the risk of harassment.

[0136] Furthermore, an emotion engine is used to analyze video data and identify the user's emotional state. This emotional information is integrated with linguistic information to enable more accurate risk assessment. This integrated system provides richer information than simple linguistic data analysis alone.

[0137] The server calculates a harassment score based on the analyzed data and determines the risk level (low, medium, or high). Based on this risk level, it generates a feedback message to notify the user. This feedback is sent to the device and provided to the user visually or audibly in real time.

[0138] For example, if a user uses aggressive language during a meeting, the device records the words and facial expressions, which are then analyzed by the server. If the emotion engine detects that the user is irritated, the server classifies it as high-risk and notifies the user via the device with feedback such as, "This statement is excessively aggressive."

[0139] The analysis data and feedback results are stored in a database. This allows the server to retrain its machine learning models based on the accumulated data, gradually improving the system's analysis accuracy. The emotion engine also continuously learns emotional patterns from past data, contributing to further improvements in the accuracy of harassment risk assessments in the future. This enables users to review their behavior more quickly and appropriately, preventing harassment proactively.

[0140] The following describes the processing flow.

[0141] Step 1:

[0142] When a user initiates a conversation, the device's microphone and camera activate, capturing audio and video data in real time. The captured data is temporarily stored within the device.

[0143] Step 2:

[0144] The device encrypts the acquired data, securely compresses it, and sends it to the server. The encrypted data is transferred through a privacy-conscious communication protocol.

[0145] Step 3:

[0146] The server converts the received audio data into text using a natural language processing (NLP) engine. This text data serves as the basis for evaluating harassment risk.

[0147] Step 4:

[0148] The server begins analyzing the text data and assesses the risk of the content of the statements by detecting specific keywords and phrase patterns.

[0149] Step 5:

[0150] The server uses video data to analyze the user's facial expressions with an emotion engine and recognize their emotional state (joy, anger, sadness, etc.). This emotion analysis allows for an understanding of the emotional context influencing their statements.

[0151] Step 6:

[0152] The server integrates the results of voice analysis and emotion analysis to calculate an overall harassment score. This score combines verbal content and emotional state, leading to a more accurate risk assessment.

[0153] Step 7:

[0154] The server classifies the risk level (low, medium, high) based on the obtained harassment score and generates a feedback message for the user. This message is adjusted to be appropriate according to the risk level.

[0155] Step 8:

[0156] A feedback message is sent from the server to the terminal, and the terminal notifies the user visually or audibly. The notification attracts the user's attention and prompts them to modify their behavior.

[0157] Step 9:

[0158] The server stores all analysis results and feedback history in a database, and uses this data to retrain the machine learning model, improving the system's analysis accuracy and the reliability of its sentiment recognition.

[0159] This processing flow allows users to receive immediate feedback on their actions, which can contribute to harassment prevention.

[0160] (Example 2)

[0161] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0162] This invention addresses the need to assess the risk of verbal and nonverbal harassment in real time in workplaces, educational institutions, and other settings, and to provide immediate feedback to users. Therefore, it requires technology that integrates audio and video analysis to capture subtle facial expressions and emotional changes that are often overlooked by conventional methods. However, current technologies have limitations in analysis accuracy and real-time capabilities, necessitating the development of systems with more advanced data analysis and feedback functions.

[0163] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0164] In this invention, the server includes means for integrating acoustic and image information to calculate a harassment score, determine a risk level, and generate a feedback message; means for notifying the user of the generated feedback message; and means for storing the analyzed information and feedback results. This enables real-time assessment of harassment risk and the provision of appropriate feedback to the user.

[0165] "Acoustic information" refers to sound data acquired using microphones and other sound acquisition devices, and includes human speech and ambient sounds.

[0166] "Image information" refers to visual data acquired using cameras and other video acquisition devices, and includes things like people's facial expressions and movements.

[0167] "Communication method" refers to protocols and means for securely sending and receiving data, including encryption technologies such as SSL / TLS.

[0168] A "natural language processing engine" is a technology that converts audio data into text, analyzes its content, and understands its meaning; it typically uses AI models.

[0169] A "facial expression recognition algorithm" is a technology that analyzes human facial expressions from image data and estimates their emotional state, utilizing deep learning models.

[0170] A "harassment score" is an index that quantitatively expresses the risk of harassment based on information obtained from audio and video.

[0171] A "feedback message" is a message containing instructions or comments that include points of criticism or warnings provided to the user based on the analyzed data.

[0172] A "computational model" is a mathematical structure that uses techniques such as machine learning to learn features from large amounts of data and perform predictions and classifications.

[0173] This invention is a system that uses acoustic and visual information to assess the risk of harassment and provides real-time feedback to the user. First, the terminal is equipped with a microphone and camera to capture the user's conversation and facial expressions. A high-sensitivity microphone and high-resolution camera are used for this purpose, enabling noise suppression and acquisition of high-definition images. The acquired data is encrypted using AES encryption technology and transmitted to the server using the SSL / TLS protocol.

[0174] The server performs a process of converting received acoustic information into text using a natural language processing engine. This process employs common speech recognition technologies. Next, an AI model is used to analyze the text data and assess risk based on specific keywords and language patterns.

[0175] Furthermore, the server analyzes the user's emotional state based on image information using facial recognition algorithms. This process utilizes libraries such as OpenCV and deep learning models to identify emotions from subtle facial expressions. The resulting emotional information and acoustic information are then integrated to assess risk in the form of a harassment score.

[0176] If a user uses offensive language during a meeting, the device records the user's facial expressions along with the words, and these are analyzed in detail on the server. If the emotion engine detects that the user is upset, the situation is assessed as high risk, and a feedback message such as "This statement is excessively offensive" is generated. This feedback is immediately communicated to the user visually or audibly through the device.

[0177] An example of a prompt message in this invention would be, "Please tell us about a situation in a workplace conversation that you found unpleasant. Please describe in detail what was said and how your feelings changed at that time." This allows users to objectively evaluate their own behavior and strive for more appropriate communication.

[0178] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0179] Step 1:

[0180] The device uses a microphone and camera to collect the user's voice and video. Audio data is acquired clearly using technology that effectively removes noise, and video data is captured in high resolution to record the user's facial expressions in detail. The input in this step is the user's raw voice and video, which are output as digital data.

[0181] Step 2:

[0182] The terminal encrypts the acquired digital data using AES encryption technology and sends it to the server using the secure protocol SSL / TLS. Security techniques are implemented during this process to ensure the safe transfer of highly confidential data. The input is the digital data before encryption, and the output is the encrypted digital data.

[0183] Step 3:

[0184] The server converts the received audio data into text data using a natural language processing engine. Specifically, it uses speech recognition technology to convert the audio signal into text information. The input is encrypted audio data, and the output is parseable text data.

[0185] Step 4:

[0186] The server performs a risk assessment using an AI model based on the converted text data. In this step, keywords and contextual patterns are analyzed to quantify the risk of harassment. This analysis process generates a risk score as output from the text data as input.

[0187] Step 5:

[0188] The server applies a facial recognition algorithm to video data to analyze the user's emotional state. Using a deep learning model, it infers emotions from subtle changes in facial expressions. The input is encrypted video data, and the output is an emotion score.

[0189] Step 6:

[0190] The server integrates data obtained from audio and video to calculate a harassment score. This integrated analysis enables more accurate harassment assessment than conventional methods. The input consists of a risk score from audio and an emotion score from images, and the output is the integrated harassment score.

[0191] Step 7:

[0192] The server assesses the risk level based on the obtained harassment score and generates a feedback message. For example, in the case of a high-risk situation, it provides appropriate feedback such as "This statement is excessively aggressive." The input is the harassment score, and the output is the generated feedback message.

[0193] Step 8:

[0194] The terminal notifies the user of the generated feedback message. Warnings and suggestions are communicated immediately using a visual display or audio output device. The input is the feedback message sent from the server, and the output is the notification to the user.

[0195] Step 9:

[0196] The server records the analyzed data and feedback history in a database. This enables long-term data analysis and contributes to improving the system's accuracy. The input is the analysis results and feedback history, and the output is the updated database.

[0197] Step 10:

[0198] The server retrains the generative AI model based on the accumulated data. This process refines the machine learning algorithm, enabling it to continue learning from new data patterns. The input is stored historical data, and the output is the improved AI model.

[0199] (Application Example 2)

[0200] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0201] In brick-and-mortar stores, there is a lack of systems to immediately detect the risk of harassment in communication between staff and customers and to enable staff to take appropriate action. This problem can lead to misunderstandings and inappropriate responses between customers, potentially resulting in a decline in service quality and unnecessary trouble. Therefore, a system that ensures safe and smooth communication is needed.

[0202] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0203] In this invention, the server includes an acquisition means for acquiring audio and video information; a transmission means for transmitting the acquired audio and video information via a communication standard; an evaluation means for analyzing the received audio information with a natural language processing device and evaluating the degree of danger of the spoken content; an analysis means for analyzing emotional states from video information using facial expression recognition calculations; a generation means for determining the degree of danger based on harassment indicators and generating feedback information; a notification means for notifying the user of the generated feedback information; a storage means for storing analysis information and history; a training means for retraining a machine learning model based on the stored information; and a notification means using a visual device to detect signs of harassment in real time and provide feedback to staff. This makes it possible to immediately assess the risk of harassment during conversations in physical stores and send feedback to staff.

[0204] "Audio and video information" refers to digital data that records the user's speech, facial expressions, and actions.

[0205] "Means of acquisition" refers to devices and technologies for collecting audio and video information.

[0206] "Communication standards" refer to protocols and specifications for sending and receiving data.

[0207] "Transmission means" refers to devices or technologies used to send acquired information to other devices such as servers.

[0208] A "natural language processing system" is a technology that analyzes speech information and understands it as text data.

[0209] "Evaluation methods" refer to devices and technologies used to analyze acquired information and determine its degree of risk.

[0210] "Facial expression recognition calculation" is an algorithm used to analyze an individual's emotions and state of mind from video information.

[0211] "Analysis means" refers to devices and technologies used to analyze acquired information in detail and extract its meaning.

[0212] "Harassment indicators" are criteria or indicators used to evaluate whether communication constitutes harassment.

[0213] "Generation means" refers to devices or technologies for forming feedback information based on analysis results.

[0214] "Notification means" refers to devices or technologies used to transmit generated feedback information to the user.

[0215] "Storage methods" refer to devices and technologies for storing analysis results and historical data.

[0216] "Training methods" refer to techniques for updating machine learning models using stored data.

[0217] A "visual device" is a device that allows users to receive feedback information visually.

[0218] The system realizing this invention acquires audio and video information in real time and evaluates the risk of harassment based on that information. Within the target physical store, staff wear smart glasses and communicate with a server via wireless communication to exchange information. The smart glasses have a built-in camera and microphone, which are used to acquire audio and video information.

[0219] The server converts audio data into text using natural language processing software (e.g., Google® Cloud Speech-to-Text) and analyzes the content of the speech. Simultaneously, video data is analyzed by facial expression recognition software (e.g., Microsoft® Azure® Emotion API) to assess the emotional state of customers and staff. These analysis results are comprehensively evaluated based on harassment indicators to determine the level of risk.

[0220] Based on the risk level, the server immediately generates feedback information and displays it on the smart glasses' display to notify staff. This visual notification provides staff with guidance for taking appropriate action. Furthermore, the analyzed information is stored in a secure database and used to retrain machine learning models. Through this process, the system improves in accuracy over time, enabling it to provide more precise feedback.

[0221] For example, if a store detects signs of customer dissatisfaction, staff can receive feedback through smart glasses such as, "The customer's tone is agitated. Please try to remain calm." This allows staff to calmly assess the situation and improve service quality.

[0222] An example of a prompt message is: "If recent customer interactions have detected voice and facial expressions indicating increased stress, what measures should be taken? What message should be displayed in the text notification?"

[0223] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0224] Step 1:

[0225] The device (smart glasses) acquires real-time audio and video information of staff and customers within the store. The input is raw data from the camera and microphone, which is then formatted as digital information. Audio files and video clips are generated as output.

[0226] Step 2:

[0227] The terminal transmits the acquired audio and video information to the server. The input is the digital information generated in step 1. The data is encrypted during transmission and securely sent to the server via communication standards. The server receives the encrypted information as output.

[0228] Step 3:

[0229] The server analyzes the received audio data using a natural language processing unit. The input is an encrypted audio file, which is decrypted and converted into text data. The output is the transcribed speech content, which is then analyzed to prepare for risk assessment.

[0230] Step 4:

[0231] The video data is analyzed on a server using facial recognition software. The input is an encrypted video clip, which is decrypted to evaluate the emotional state of staff and customers. The output is the analyzed emotional data, which is used for risk assessment.

[0232] Step 5:

[0233] The server integrates transcribed speech content and sentiment data, and evaluates the risk level based on harassment indicators. The input is the analysis results from steps 3 and 4. The data is integrated to determine the risk level (low, medium, high). The output is the risk level and the feedback information that should be taken.

[0234] Step 6:

[0235] The server generates feedback information and notifies the terminal. The input is the feedback information determined in step 5. This is sent to the terminal's display to provide visual feedback to the staff. The output is a feedback display on the terminal's display that is recognizable to the staff.

[0236] Step 7:

[0237] The server encrypts the analysis data and risk assessment history and stores it in a database. The input is the assessment results and feedback history from step 5. This is saved and used as training data for future machine learning models. The output is the updated database.

[0238] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0239] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0240] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0241] [Second Embodiment]

[0242] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0243] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0244] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0245] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0246] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0247] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0248] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0249] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0250] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0251] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0252] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0253] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. 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".

[0254] The system of the present invention acquires audio and video data in real time, analyzes it to assess the risk of harassment, and provides feedback to the user.

[0255] In this embodiment, the terminal first uses a microphone and camera to acquire the user's voice and video as data. This data is transmitted to the server in an encrypted format. To protect privacy, advanced encryption protocols are used for communication during data transmission.

[0256] The server analyzes the received data. First, it converts audio data into text using natural language processing (NLP) technology, and then measures the likelihood of harassment based on the text structure and content. Simultaneously, it analyzes the user's emotional state using facial recognition algorithms based on video data. This analysis allows for a multifaceted evaluation of the user's intentions and how they were perceived.

[0257] Next, the server scores the harassment risk based on the analysis results and determines the risk level (low, medium, or high) based on that score. It then generates an appropriate feedback message accordingly. This generated feedback is returned to the terminal in real time and notified to the user. Based on the feedback, the user can immediately review their own words and actions.

[0258] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and facial expressions and sends them to the server. The server assigns a high risk score to this comment, citing gender bias, and notifies the user through the device that "this comment may be inappropriate."

[0259] Furthermore, the server stores a history of all analysis and feedback, and uses this to retrain the machine learning model, improving accuracy and reliability. This process allows the system to continuously provide intuitive and useful feedback to users, thereby preventing harassment and improving the quality of communication.

[0260] The following describes the processing flow.

[0261] Step 1:

[0262] The device uses a microphone and camera to acquire the user's voice and video data. The data is captured in real time and temporarily stored by the system.

[0263] Step 2:

[0264] The device compresses the acquired audio and video data and sends it to the server in a highly encrypted format. This enhances data security.

[0265] Step 3:

[0266] The server uses a natural language processing (NLP) engine to convert the received audio data into text. Next, it identifies specific keywords and language patterns from the analyzed text to assess the potential for harassment.

[0267] Step 4:

[0268] The server analyzes the received video data using a facial recognition algorithm to evaluate the user's emotional state. This information is combined with text analysis results to perform a comprehensive risk assessment.

[0269] Step 5:

[0270] The server analyzes audio and video data to score the harassment risk and classifies it into three risk levels (low, medium, high). It then generates a feedback message corresponding to this risk level.

[0271] Step 6:

[0272] The server sends the generated feedback message to the terminal.

[0273] Step 7:

[0274] The device notifies the user of the generated feedback message in real time. This gives the user an opportunity to review their own words and actions.

[0275] Step 8:

[0276] The server stores the analysis results and user feedback history in a database.

[0277] Step 9:

[0278] The server uses the stored data to retrain the machine learning model and improve the system's accuracy. This process is performed periodically.

[0279] (Example 1)

[0280] Next, we will describe Example 1. 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".

[0281] In today's communication environment, it is essential to recognize the risks of harassment early and for users to appropriately avoid them. However, conventional technologies do not adequately provide efficient methods for detecting harassment using audio and video, and there is a lack of mechanisms to provide real-time feedback on inappropriate remarks and actions that users may unconsciously make. As a result, the detection and correction of harassment is delayed, hindering the improvement of the quality of communication.

[0282] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0283] In this invention, the server includes acquisition means for acquiring acoustic and image data, transmission means for transmitting this data via an information transmission protocol, and analysis means that use a natural language processing engine and image recognition algorithms to analyze the received data, assess the risks, and provide feedback. This enables early detection of harassment risks, and allows users to receive real-time feedback, enabling them to consciously and quickly review their behavior.

[0284] "Acoustic data" refers to data in digital or analog form that contains voice information and is acquired for the purpose of information transmission based on sound waves.

[0285] "Image data" refers to data in digital or analog form that contains visual information acquired by a camera or other imaging device.

[0286] "Information transmission protocol" refers to a set of rules and procedures for securely and efficiently communicating data.

[0287] "Natural language processing engine" refers to a set of algorithms for analyzing voice or text data to assist in the understanding and generation of natural languages used by humans.

[0288] "Image recognition algorithm" refers to an algorithm for analyzing specific patterns and features in an image to extract and recognize information.

[0289] "Harassment evaluation criteria" refers to criteria or standards used to quantify the likelihood and degree of harassment.

[0290] "Machine learning model" refers to a framework of algorithms for learning patterns from data and making predictions or classifications.

[0291] As an embodiment of the present invention, first, the terminal uses a microphone and a camera to acquire acoustic data and image data. This includes electronic devices, voice recognition microphones, and HD cameras commonly used in daily life. This information is encrypted in real-time and transmitted to the server via an information transmission protocol.

[0292] Next, the server converts the received audio data into text data using a natural language processing engine. This engine includes, for example, speech recognition software. The server then analyzes the text data and assesses the potential risks based on the content of the speech. Simultaneously, the server analyzes the image data using an image recognition algorithm to determine the user's emotional state. This image recognition algorithm utilizes facial recognition software.

[0293] Based on these analysis results, the server generates harassment evaluation metrics and calculates a risk score in real time. If a high risk is determined, the AI ​​model generates an appropriate feedback message for the user and notifies them through their device.

[0294] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and video information and sends it to the server. The server determines that the comment contains gender bias and scores it as high risk. It then sends feedback to the device stating, "This comment may be inappropriate," and notifies the user. This notification allows the user to immediately review their comment.

[0295] An example of a prompt message would be: "Voice: This job is probably impossible for a woman. Emotional state: Neutral. Please analyze."

[0296] This system allows users to recognize the risk of harassment they may have unintentionally caused early on and have the opportunity to improve it. Furthermore, the server stores all analysis results and feedback history, and uses this data to retrain the computer learning model, continuously improving the overall accuracy and reliability of the system.

[0297] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0298] Step 1:

[0299] The terminal acquires acoustic data and image data from the user. It records sound with a microphone and takes pictures with a camera. The acquired data serves as the input. The data is encrypted and transmitted to the server using an information transfer protocol. The output is the encrypted acoustic data and image data.

[0300] Step 2:

[0301] The server inputs the received acoustic data into a natural language processing engine and converts it into text data. Here, speech recognition technology is used to analyze the language from the acoustic waveform and form sentences. Through this process, the content of the speech is output in text form.

[0302] Step 3:

[0303] At the same time, the server uses an image recognition algorithm to input the image data and analyze the user's emotional state. In the process of analysis, it captures the feature points of the expression, discriminates subtle expression changes, and outputs the emotional state (such as joy, anger, sadness, etc.).

[0304] Step 4:

[0305] The server integrates the text data and the results of the image emotional analysis, and calculates the harassment evaluation index based on these. Specifically, it utilizes a generative AI model to input the text and emotional information, and scores the risk of the speech. The output is a numerical score of the harassment risk.

[0306] Step 5:

[0307] Based on the calculated risk score, the server generates feedback information. It uses a generative AI model to create a text message according to the risk level, and the content of the message is output.

[0308] Step 6:

[0309] The server sends the generated feedback information to the terminal. The terminal notifies the user of the received feedback, allowing the user to modify their actions based on it. Real-time notifications to the user are output.

[0310] Step 7:

[0311] The server stores all analysis results and feedback history. This dataset is used to retrain the computer learning model, continuously improving the system's accuracy and reliability. Data storage and model training are the outputs.

[0312] (Application Example 1)

[0313] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0314] In family communication, unintentional harassment and inappropriate remarks can damage relationships with family members and guests. Preventing such problems and maintaining a smooth and harmonious communication environment is crucial. However, currently, there is a lack of means to detect these problems in real time and provide feedback. Therefore, there is a need for a system that can address this challenge.

[0315] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0316] In this invention, the server includes acquisition means for acquiring acoustic and visual information; transmission means for transmitting the acquired acoustic and visual information via a communication protocol; evaluation means for analyzing the received acoustic information with a natural language processing engine and evaluating the risk of the content of the speech; analysis means for analyzing emotional states from visual information using an facial expression recognition algorithm; generation means for determining the degree of risk based on harassment indicators and generating feedback information; notification means for notifying the user of the generated feedback information; storage means for storing analysis information and history information; training means for retraining a machine learning model based on the stored information; and means having a function to monitor conversations within the home, detect specific phrases to support smooth communication, and provide feedback. This makes it possible to improve the quality of communication within the home.

[0317] "Acoustic information" refers to data used to acquire ambient sounds and speech.

[0318] "Visual information" refers to video data acquired through a camera.

[0319] A "communication protocol" is a set of rules that define how data can be transmitted securely.

[0320] A "natural language processing engine" is a technology that converts audio data into text and analyzes its content.

[0321] A "facial expression recognition algorithm" is a method for analyzing a user's emotional state from video data.

[0322] A "harassment index" is a set of values ​​used as a standard for evaluating the risks of words and actions.

[0323] "Feedback information" refers to advice and warning messages provided to users based on the analysis results.

[0324] A "machine learning model" is a collection of algorithms used to perform analysis and predictions based on data.

[0325] "Monitoring conversations within the household" means monitoring the speech and actions of members of the household to detect potentially dangerous behaviors.

[0326] This invention is a system designed to improve the quality of communication within the home. To acquire acoustic and visual information, the server uses a terminal equipped with a high-performance microphone and camera. This allows for the acquisition of acoustic and video data in real time. This data is transmitted to the server via a communication protocol. The communication protocol uses encryption technologies such as TLS / SSL to ensure secure data transmission.

[0327] On the server, received audio data is converted into text using a natural language processing engine (e.g., the spaCy library) and the potential danger of the spoken content is evaluated. For visual information, the user's emotional state is analyzed using an facial recognition algorithm based on OpenCV. Based on these analysis results, a harassment index is calculated and the degree of danger is evaluated.

[0328] Based on the assessed risk level, the server generates feedback information. This generated feedback information is then notified to the user. Since the feedback information is presented in real time through the device's display or speaker, the user can immediately review their own statements and actions.

[0329] Furthermore, the server stores analysis information and feedback history. Based on this stored information, the machine learning model can be retrained to improve the accuracy and reliability of the feedback. By detecting specific phrases in conversations within the home and providing appropriate feedback, it supports smoother communication.

[0330] For example, if someone says "My opinion might be considered boring" during a family dinner, the server can pick up on this phrase, analyze the emotion behind it, and then provide feedback such as, "Why not try trusting yourself more and sharing your opinion?"

[0331] An example of a prompt to input into a generative AI model is, "Please describe how to assess the harassment risk in conversations within the family and provide feedback on specific areas for improvement."

[0332] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0333] Step 1:

[0334] The device uses a high-performance microphone and camera to acquire acoustic and visual information within the home. This input consists of audio and video data collected in real time. The device stores this data in a buffer and prepares it for subsequent processing.

[0335] Step 2:

[0336] The terminal transmits the acquired audio and visual information to the server via a secure communication protocol using encryption technologies such as TLS / SSL. The input data is raw audio and video, and the output is this data encrypted and sent to the server. The server decrypts the received data and makes it analyzable.

[0337] Step 3:

[0338] The server converts the received audio data into text using a natural language processing engine (e.g., spaCy). The input for this step is audio data, and the output is the generated text data. The server then performs speech recognition and processes the conversation content as textual information.

[0339] Step 4:

[0340] The server processes the received visual information using an OpenCV-based facial recognition algorithm to analyze the user's emotional state. The input for this step is video data, and the output is numerical data representing emotions. The server analyzes the user's facial expressions and detects emotional patterns.

[0341] Step 5:

[0342] The server integrates natural language processing results and facial recognition results, and uses this to evaluate the potential danger of the spoken content. The input is text data and emotion data, and the output is a danger score as a harassment index. The server analyzes the fused data to score the potential harassment risk.

[0343] Step 6:

[0344] The server generates feedback information based on the risk score and sends it to the terminal. The input is the risk score, and the output is a feedback message to notify the user. The server then forms an appropriate advice or warning message.

[0345] Step 7:

[0346] The terminal notifies the user of feedback information received from the server via its display or speaker. The input is the feedback message, and the output is the transmission of information visually or audibly. This allows the user to immediately review their own actions.

[0347] Step 8:

[0348] The server stores all analysis results and feedback history in a database. The inputs are the analysis data and feedback history, and the output is the recorded data. The server stores this data for future retraining.

[0349] Step 9:

[0350] The server retrains a machine learning model based on stored data. The input is the stored historical data, and the output is the improved model. The server periodically updates the model to improve analysis accuracy and increase the reliability of the feedback.

[0351] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0352] This invention is a system that acquires user voice and video data, assesses the risk of harassment based on this data, and provides immediate feedback to the user. The system further incorporates an emotion engine to recognize the user's emotions.

[0353] First, the device uses its microphone and camera to capture data on the user's conversation and facial expressions. This data is securely processed using encryption technology and sent to the server. The transmitted data is guaranteed to be encrypted to ensure the user's privacy.

[0354] Next, the server converts the received audio data into text using natural language processing (NLP) technology. Based on this text information, it analyzes specific keywords and language patterns to assess the risk of harassment.

[0355] Furthermore, an emotion engine is used to analyze video data and identify the user's emotional state. This emotional information is integrated with linguistic information to enable more accurate risk assessment. This integrated system provides richer information than simple linguistic data analysis alone.

[0356] The server calculates a harassment score based on the analyzed data and determines the risk level (low, medium, or high). Based on this risk level, it generates a feedback message to notify the user. This feedback is sent to the device and provided to the user visually or audibly in real time.

[0357] For example, if a user uses aggressive language during a meeting, the device records the words and facial expressions, which are then analyzed by the server. If the emotion engine detects that the user is irritated, the server classifies it as high-risk and notifies the user via the device with feedback such as, "This statement is excessively aggressive."

[0358] The analysis data and feedback results are stored in a database. This allows the server to retrain its machine learning models based on the accumulated data, gradually improving the system's analysis accuracy. The emotion engine also continuously learns emotional patterns from past data, contributing to further improvements in the accuracy of harassment risk assessments in the future. This enables users to review their behavior more quickly and appropriately, preventing harassment proactively.

[0359] The following describes the processing flow.

[0360] Step 1:

[0361] When a user initiates a conversation, the device's microphone and camera activate, capturing audio and video data in real time. The captured data is temporarily stored within the device.

[0362] Step 2:

[0363] The device encrypts the acquired data, securely compresses it, and sends it to the server. The encrypted data is transferred through a privacy-conscious communication protocol.

[0364] Step 3:

[0365] The server converts the received audio data into text using a natural language processing (NLP) engine. This text data serves as the basis for evaluating harassment risk.

[0366] Step 4:

[0367] The server begins analyzing the text data and assesses the risk of the content of the statements by detecting specific keywords and phrase patterns.

[0368] Step 5:

[0369] The server uses video data to analyze the user's facial expressions with an emotion engine and recognize their emotional state (joy, anger, sadness, etc.). This emotion analysis allows for an understanding of the emotional context influencing their statements.

[0370] Step 6:

[0371] The server integrates the results of voice analysis and emotion analysis to calculate an overall harassment score. This score combines verbal content and emotional state, leading to a more accurate risk assessment.

[0372] Step 7:

[0373] The server classifies the risk level (low, medium, high) based on the obtained harassment score and generates a feedback message for the user. This message is adjusted to be appropriate according to the risk level.

[0374] Step 8:

[0375] A feedback message is sent from the server to the terminal, and the terminal notifies the user visually or audibly. The notification attracts the user's attention and prompts them to modify their behavior.

[0376] Step 9:

[0377] The server stores all analysis results and feedback history in a database, and uses this data to retrain the machine learning model, improving the system's analysis accuracy and the reliability of its sentiment recognition.

[0378] This processing flow allows users to receive immediate feedback on their actions, which can contribute to harassment prevention.

[0379] (Example 2)

[0380] Next, we will describe Example 2. 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".

[0381] This invention addresses the need to assess the risk of verbal and nonverbal harassment in real time in workplaces, educational institutions, and other settings, and to provide immediate feedback to users. Therefore, it requires technology that integrates audio and video analysis to capture subtle facial expressions and emotional changes that are often overlooked by conventional methods. However, current technologies have limitations in analysis accuracy and real-time capabilities, necessitating the development of systems with more advanced data analysis and feedback functions.

[0382] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0383] In this invention, the server includes means for integrating acoustic and image information to calculate a harassment score, determine a risk level, and generate a feedback message; means for notifying the user of the generated feedback message; and means for storing the analyzed information and feedback results. This enables real-time assessment of harassment risk and the provision of appropriate feedback to the user.

[0384] "Acoustic information" refers to sound data acquired using microphones and other sound acquisition devices, and includes human speech and ambient sounds.

[0385] "Image information" refers to visual data acquired using cameras and other video acquisition devices, and includes things like people's facial expressions and movements.

[0386] "Communication method" refers to protocols and means for securely sending and receiving data, including encryption technologies such as SSL / TLS.

[0387] A "natural language processing engine" is a technology that converts audio data into text, analyzes its content, and understands its meaning; it typically uses AI models.

[0388] A "facial expression recognition algorithm" is a technology that analyzes human facial expressions from image data and estimates their emotional state, utilizing deep learning models.

[0389] A "harassment score" is an index that quantitatively expresses the risk of harassment based on information obtained from audio and video.

[0390] A "feedback message" is a message containing instructions or comments, including points of criticism or warnings, provided to the user based on the analyzed data.

[0391] A "computational model" is a mathematical structure that uses techniques such as machine learning to learn features from large amounts of data and perform predictions and classifications.

[0392] This invention is a system that uses acoustic and visual information to assess the risk of harassment and provides real-time feedback to the user. First, the terminal is equipped with a microphone and camera to capture the user's conversation and facial expressions. A high-sensitivity microphone and high-resolution camera are used for this purpose, enabling noise suppression and acquisition of high-definition images. The acquired data is encrypted using AES encryption technology and transmitted to the server using the SSL / TLS protocol.

[0393] The server performs a process of converting received acoustic information into text using a natural language processing engine. This process employs common speech recognition technologies. Next, an AI model is used to analyze the text data and assess risk based on specific keywords and language patterns.

[0394] Furthermore, the server analyzes the user's emotional state based on image information using facial recognition algorithms. This process utilizes libraries such as OpenCV and deep learning models to identify emotions from subtle facial expressions. The resulting emotional information and acoustic information are then integrated to assess risk in the form of a harassment score.

[0395] If a user uses offensive language during a meeting, the device records the user's facial expressions along with the words, and these are analyzed in detail on the server. If the emotion engine detects that the user is upset, the situation is assessed as high risk, and a feedback message such as "This statement is excessively offensive" is generated. This feedback is immediately communicated to the user visually or audibly through the device.

[0396] An example of a prompt message in this invention would be, "Please tell us about a situation in a workplace conversation that you found unpleasant. Please describe in detail what was said and how your feelings changed at that time." This allows users to objectively evaluate their own behavior and strive for more appropriate communication.

[0397] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0398] Step 1:

[0399] The device uses a microphone and camera to collect the user's voice and video. Audio data is acquired clearly using technology that effectively removes noise, and video data is captured in high resolution to record the user's facial expressions in detail. The input in this step is the user's raw voice and video, which are output as digital data.

[0400] Step 2:

[0401] The terminal encrypts the acquired digital data using AES encryption technology and sends it to the server using the secure protocol SSL / TLS. Security techniques are implemented during this process to ensure the safe transfer of highly confidential data. The input is the digital data before encryption, and the output is the encrypted digital data.

[0402] Step 3:

[0403] The server converts the received audio data into text data using a natural language processing engine. Specifically, it uses speech recognition technology to convert the audio signal into text information. The input is encrypted audio data, and the output is parseable text data.

[0404] Step 4:

[0405] The server performs a risk assessment using an AI model based on the converted text data. In this step, keywords and contextual patterns are analyzed to quantify the risk of harassment. This analysis process generates a risk score as output from the text data as input.

[0406] Step 5:

[0407] The server applies a facial recognition algorithm to video data to analyze the user's emotional state. Using a deep learning model, it infers emotions from subtle changes in facial expressions. The input is encrypted video data, and the output is an emotion score.

[0408] Step 6:

[0409] The server integrates data obtained from audio and video to calculate a harassment score. This integrated analysis enables more accurate harassment assessment than conventional methods. The input consists of a risk score from audio and an emotion score from images, and the output is the integrated harassment score.

[0410] Step 7:

[0411] The server assesses the risk level based on the obtained harassment score and generates a feedback message. For example, in the case of a high-risk situation, it provides appropriate feedback such as "This statement is excessively aggressive." The input is the harassment score, and the output is the generated feedback message.

[0412] Step 8:

[0413] The terminal notifies the user of the generated feedback message. Warnings and suggestions are communicated immediately using a visual display or audio output device. The input is the feedback message sent from the server, and the output is the notification to the user.

[0414] Step 9:

[0415] The server records the analyzed data and feedback history in a database. This enables long-term data analysis and contributes to improving the system's accuracy. The input is the analysis results and feedback history, and the output is the updated database.

[0416] Step 10:

[0417] The server retrains the generative AI model based on the accumulated data. This process refines the machine learning algorithm, enabling it to continue learning from new data patterns. The input is stored historical data, and the output is the improved AI model.

[0418] (Application Example 2)

[0419] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the smart glasses 214 as the "terminal".

[0420] In brick-and-mortar stores, there is a lack of systems to immediately detect the risk of harassment in communication between staff and customers and to enable staff to take appropriate action. This problem can lead to misunderstandings and inappropriate responses between customers, potentially resulting in a decline in service quality and unnecessary trouble. Therefore, a system that ensures safe and smooth communication is needed.

[0421] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0422] In this invention, the server includes an acquisition means for acquiring audio and video information; a transmission means for transmitting the acquired audio and video information via a communication standard; an evaluation means for analyzing the received audio information with a natural language processing device and evaluating the degree of danger of the spoken content; an analysis means for analyzing emotional states from video information using facial expression recognition calculations; a generation means for determining the degree of danger based on harassment indicators and generating feedback information; a notification means for notifying the user of the generated feedback information; a storage means for storing analysis information and history; a training means for retraining a machine learning model based on the stored information; and a notification means using a visual device to detect signs of harassment in real time and provide feedback to staff. This makes it possible to immediately assess the risk of harassment during conversations in physical stores and send feedback to staff.

[0423] "Audio and video information" refers to digital data that records the user's speech, facial expressions, and actions.

[0424] "Means of acquisition" refers to devices and technologies for collecting audio and video information.

[0425] "Communication standards" refer to protocols and specifications for sending and receiving data.

[0426] "Transmission means" refers to devices or technologies used to send acquired information to other devices such as servers.

[0427] A "natural language processing system" is a technology that analyzes speech information and understands it as text data.

[0428] "Evaluation methods" refer to devices and technologies used to analyze acquired information and determine its degree of risk.

[0429] "Facial expression recognition calculation" is an algorithm used to analyze an individual's emotions and state of mind from video information.

[0430] "Analysis means" refers to devices and technologies used to analyze acquired information in detail and extract its meaning.

[0431] "Harassment indicators" are criteria or indicators used to evaluate whether communication constitutes harassment.

[0432] "Generation means" refers to devices or technologies for forming feedback information based on analysis results.

[0433] "Notification means" refers to devices or technologies used to transmit generated feedback information to the user.

[0434] "Storage methods" refer to devices and technologies for storing analysis results and historical data.

[0435] "Training methods" refer to techniques for updating machine learning models using stored data.

[0436] A "visual device" is a device that allows users to receive feedback information visually.

[0437] The system realizing this invention acquires audio and video information in real time and evaluates the risk of harassment based on that information. Within the target physical store, staff wear smart glasses and communicate with a server via wireless communication to exchange information. The smart glasses have a built-in camera and microphone, which are used to acquire audio and video information.

[0438] The server converts audio data into text using a natural language processing unit (e.g., Google Cloud Speech-to-Text) and analyzes the content of the speech. Simultaneously, video data is analyzed by facial expression recognition software (e.g., Microsoft Azure Emotion API) to assess the emotional state of customers and staff. These analysis results are comprehensively evaluated based on harassment indicators to determine the level of risk.

[0439] Based on the risk level, the server immediately generates feedback information and displays it on the smart glasses' display to notify staff. This visual notification provides staff with guidance for taking appropriate action. Furthermore, the analyzed information is stored in a secure database and used to retrain machine learning models. Through this process, the system improves in accuracy over time, enabling it to provide more precise feedback.

[0440] For example, if a store detects signs of customer dissatisfaction, staff can receive feedback through smart glasses such as, "The customer's tone is agitated. Please try to remain calm." This allows staff to calmly assess the situation and improve service quality.

[0441] An example of a prompt message is: "If recent customer interactions have detected voice and facial expressions indicating increased stress, what measures should be taken? What message should be displayed in the text notification?"

[0442] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0443] Step 1:

[0444] The device (smart glasses) acquires real-time audio and video information of staff and customers within the store. The input is raw data from the camera and microphone, which is then formatted as digital information. Audio files and video clips are generated as output.

[0445] Step 2:

[0446] The terminal transmits the acquired audio and video information to the server. The input is the digital information generated in step 1. The data is encrypted during transmission and securely sent to the server via communication standards. The server receives the encrypted information as output.

[0447] Step 3:

[0448] The server analyzes the received audio data using a natural language processing unit. The input is an encrypted audio file, which is decrypted and converted into text data. The output is the transcribed speech content, which is then analyzed to prepare for risk assessment.

[0449] Step 4:

[0450] The video data is analyzed on a server using facial recognition software. The input is an encrypted video clip, which is decrypted to evaluate the emotional state of staff and customers. The output is the analyzed emotional data, which is used for risk assessment.

[0451] Step 5:

[0452] The server integrates transcribed speech content and sentiment data, and evaluates the risk level based on harassment indicators. The input is the analysis results from steps 3 and 4. The data is integrated to determine the risk level (low, medium, high). The output is the risk level and the feedback information that should be taken.

[0453] Step 6:

[0454] The server generates feedback information and notifies the terminal. The input is the feedback information determined in step 5. This is sent to the terminal's display to provide visual feedback to the staff. The output is a feedback display on the terminal's display that is recognizable to the staff.

[0455] Step 7:

[0456] The server encrypts the analysis data and risk assessment history and stores it in a database. The input is the assessment results and feedback history from step 5. This is saved and used as training data for future machine learning models. The output is the updated database.

[0457] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0458] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0459] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0460] [Third Embodiment]

[0461] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0462] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0463] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0464] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0465] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0466] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0467] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0468] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0469] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0470] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0471] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0472] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0473] The system of the present invention acquires audio and video data in real time, analyzes it to assess the risk of harassment, and provides feedback to the user.

[0474] In this embodiment, the terminal first uses a microphone and camera to acquire the user's voice and video as data. This data is transmitted to the server in an encrypted format. To protect privacy, advanced encryption protocols are used for communication during data transmission.

[0475] The server analyzes the received data. First, it converts audio data into text using natural language processing (NLP) technology, and then measures the likelihood of harassment based on the text structure and content. Simultaneously, it analyzes the user's emotional state using facial recognition algorithms based on video data. This analysis allows for a multifaceted evaluation of the user's intentions and how they were perceived.

[0476] Next, the server scores the harassment risk based on the analysis results and determines the risk level (low, medium, or high) based on that score. It then generates an appropriate feedback message accordingly. This generated feedback is returned to the terminal in real time and notified to the user. Based on the feedback, the user can immediately review their own words and actions.

[0477] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and facial expressions and sends them to the server. The server assigns a high risk score to this comment, citing gender bias, and notifies the user through the device that "this comment may be inappropriate."

[0478] Furthermore, the server stores a history of all analysis and feedback, and uses this to retrain the machine learning model, improving accuracy and reliability. This process allows the system to continuously provide intuitive and useful feedback to users, thereby preventing harassment and improving the quality of communication.

[0479] The following describes the processing flow.

[0480] Step 1:

[0481] The device uses a microphone and camera to acquire the user's voice and video data. The data is captured in real time and temporarily stored by the system.

[0482] Step 2:

[0483] The device compresses the acquired audio and video data and sends it to the server in a highly encrypted format. This enhances data security.

[0484] Step 3:

[0485] The server uses a natural language processing (NLP) engine to convert the received audio data into text. Next, it identifies specific keywords and language patterns from the analyzed text to assess the potential for harassment.

[0486] Step 4:

[0487] The server analyzes the received video data using a facial recognition algorithm to evaluate the user's emotional state. This information is combined with text analysis results to perform a comprehensive risk assessment.

[0488] Step 5:

[0489] The server analyzes audio and video data to score the harassment risk and classifies it into three risk levels (low, medium, high). It then generates a feedback message corresponding to this risk level.

[0490] Step 6:

[0491] The server sends the generated feedback message to the terminal.

[0492] Step 7:

[0493] The device notifies the user of the generated feedback message in real time. This gives the user an opportunity to review their own words and actions.

[0494] Step 8:

[0495] The server stores the analysis results and user feedback history in a database.

[0496] Step 9:

[0497] The server uses the stored data to retrain the machine learning model and improve the system's accuracy. This process is performed periodically.

[0498] (Example 1)

[0499] Next, we will describe Example 1. 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."

[0500] In today's communication environment, it is essential to recognize the risks of harassment early and for users to appropriately avoid them. However, conventional technologies do not adequately provide efficient methods for detecting harassment using audio and video, and there is a lack of mechanisms to provide real-time feedback on inappropriate remarks and actions that users may unconsciously make. As a result, the detection and correction of harassment is delayed, hindering the improvement of the quality of communication.

[0501] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0502] In this invention, the server includes acquisition means for acquiring acoustic and image data, transmission means for transmitting this data via an information transmission protocol, and analysis means that use a natural language processing engine and image recognition algorithms to analyze the received data, assess the risks, and provide feedback. This enables early detection of harassment risks, and allows users to receive real-time feedback, enabling them to consciously and quickly review their behavior.

[0503] "Audio data" refers to digital or analog data containing audio information, acquired for the purpose of transmitting information based on sound waves.

[0504] "Image data" refers to digital or analog data containing visual information acquired by cameras or other imaging devices.

[0505] An "information transmission protocol" is a set of rules and procedures for communicating data securely and efficiently.

[0506] A "natural language processing engine" is a set of algorithms that analyze speech or text data to support the understanding and generation of natural language used by humans.

[0507] An "image recognition algorithm" is an algorithm that analyzes specific patterns or features within an image to extract and recognize information.

[0508] A "harassment evaluation index" is a standard or criterion used to quantify the possibility and degree of harassment.

[0509] A "computer learning model" is a framework of algorithms that learns patterns from data and performs predictions and classifications.

[0510] In an embodiment of the present invention, a terminal first acquires acoustic and image data using a microphone and a camera. This includes commonly used electronic devices, voice recognition microphones, and HD cameras. This information is encrypted in real time and transmitted to a server via an information transmission protocol.

[0511] Next, the server converts the received audio data into text data using a natural language processing engine. This engine includes, for example, speech recognition software. The server then analyzes the text data and assesses the potential risks based on the content of the speech. Simultaneously, the server analyzes the image data using an image recognition algorithm to determine the user's emotional state. This image recognition algorithm utilizes facial recognition software.

[0512] Based on these analysis results, the server generates harassment evaluation metrics and calculates a risk score in real time. If a high risk is determined, the AI ​​model generates an appropriate feedback message for the user and notifies them through their device.

[0513] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and video information and sends it to the server. The server determines that the comment contains gender bias and scores it as high risk. It then sends feedback to the device stating, "This comment may be inappropriate," and notifies the user. This notification allows the user to immediately review their comment.

[0514] An example of a prompt message would be: "Voice: This job is probably impossible for a woman. Emotional state: Neutral. Please analyze."

[0515] This system allows users to recognize the risk of harassment they may have unintentionally caused early on and have the opportunity to improve it. Furthermore, the server stores all analysis results and feedback history, and uses this data to retrain the computer learning model, continuously improving the overall accuracy and reliability of the system.

[0516] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0517] Step 1:

[0518] The device acquires audio and image data from the user. It records audio with a microphone and captures images with a camera. This acquired data serves as input. The data is encrypted and sent to the server using an information transmission protocol. The output consists of encrypted audio and image data.

[0519] Step 2:

[0520] The server inputs the received audio data into a natural language processing engine, which converts it into text data. Here, speech recognition technology is used to analyze the language from the audio waveform and transcribe it into text. This process outputs the audio content in text format.

[0521] Step 3:

[0522] Simultaneously, the server uses an image recognition algorithm to input image data and analyze the user's emotional state. During the analysis process, it captures feature points of facial expressions, identifies subtle changes in facial expression, and outputs the emotional state (e.g., joy, anger, sadness).

[0523] Step 4:

[0524] The server integrates text data and image sentiment analysis results, and uses this information to calculate a harassment assessment index. Specifically, it utilizes a generative AI model to input text and sentiment information and score the risk level of the remarks. The output is a numerical score of harassment risk.

[0525] Step 5:

[0526] The server generates feedback information based on the calculated risk score. Using a generation AI model, it creates text messages corresponding to the risk level, and the content of those messages is output.

[0527] Step 6:

[0528] The server sends the generated feedback information to the terminal. The terminal notifies the user of the received feedback, allowing the user to modify their actions based on it. Real-time notifications to the user are output.

[0529] Step 7:

[0530] The server stores all analysis results and feedback history. This dataset is used to retrain the computer learning model, continuously improving the system's accuracy and reliability. Data storage and model training are the outputs.

[0531] (Application Example 1)

[0532] Next, we will explain Application Example 1. In the following explanation, 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."

[0533] In family communication, unintentional harassment and inappropriate remarks can damage relationships with family members and guests. Preventing such problems and maintaining a smooth and harmonious communication environment is crucial. However, currently, there is a lack of means to detect these problems in real time and provide feedback. Therefore, there is a need for a system that can address this challenge.

[0534] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0535] In this invention, the server includes acquisition means for acquiring acoustic and visual information; transmission means for transmitting the acquired acoustic and visual information via a communication protocol; evaluation means for analyzing the received acoustic information with a natural language processing engine and evaluating the risk of the content of the speech; analysis means for analyzing emotional states from visual information using an facial expression recognition algorithm; generation means for determining the degree of risk based on harassment indicators and generating feedback information; notification means for notifying the user of the generated feedback information; storage means for storing analysis information and history information; training means for retraining a machine learning model based on the stored information; and means having a function to monitor conversations within the home, detect specific phrases to support smooth communication, and provide feedback. This makes it possible to improve the quality of communication within the home.

[0536] "Acoustic information" refers to data used to acquire ambient sounds and speech.

[0537] "Visual information" refers to video data acquired through a camera.

[0538] A "communication protocol" is a set of rules that define how data can be transmitted securely.

[0539] A "natural language processing engine" is a technology that converts audio data into text and analyzes its content.

[0540] A "facial expression recognition algorithm" is a method for analyzing a user's emotional state from video data.

[0541] A "harassment index" is a set of values ​​used as a standard for evaluating the risks of words and actions.

[0542] "Feedback information" refers to advice and warning messages provided to users based on the analysis results.

[0543] A "machine learning model" is a collection of algorithms used to perform analysis and predictions based on data.

[0544] "Monitoring conversations within the household" means monitoring the speech and actions of members of the household to detect potentially dangerous behaviors.

[0545] This invention is a system designed to improve the quality of communication within the home. To acquire acoustic and visual information, the server uses a terminal equipped with a high-performance microphone and camera. This allows for the acquisition of acoustic and video data in real time. This data is transmitted to the server via a communication protocol. The communication protocol uses encryption technologies such as TLS / SSL to ensure secure data transmission.

[0546] On the server, received audio data is converted into text using a natural language processing engine (e.g., the spaCy library) and the potential danger of the spoken content is evaluated. For visual information, the user's emotional state is analyzed using an facial recognition algorithm based on OpenCV. Based on these analysis results, a harassment index is calculated and the degree of danger is evaluated.

[0547] Based on the assessed risk level, the server generates feedback information. This generated feedback information is then notified to the user. Since the feedback information is presented in real time through the device's display or speaker, the user can immediately review their own statements and actions.

[0548] Furthermore, the server stores analysis information and feedback history. Based on this stored information, the machine learning model can be retrained to improve the accuracy and reliability of the feedback. By detecting specific phrases in conversations within the home and providing appropriate feedback, it supports smoother communication.

[0549] For example, if someone says "My opinion might be considered boring" during a family dinner, the server can pick up on this phrase, analyze the emotion behind it, and then provide feedback such as, "Why not try trusting yourself more and sharing your opinion?"

[0550] An example of a prompt to input into a generative AI model is, "Please describe how to assess the harassment risk in conversations within the family and provide feedback on specific areas for improvement."

[0551] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0552] Step 1:

[0553] The device uses a high-performance microphone and camera to acquire acoustic and visual information within the home. This input consists of audio and video data collected in real time. The device stores this data in a buffer and prepares it for subsequent processing.

[0554] Step 2:

[0555] The terminal transmits the acquired audio and visual information to the server via a secure communication protocol using encryption technologies such as TLS / SSL. The input data is raw audio and video, and the output is this data encrypted and sent to the server. The server decrypts the received data and makes it analyzable.

[0556] Step 3:

[0557] The server converts the received audio data into text using a natural language processing engine (e.g., spaCy). The input for this step is audio data, and the output is the generated text data. The server then performs speech recognition and processes the conversation content as textual information.

[0558] Step 4:

[0559] The server processes the received visual information using an OpenCV-based facial recognition algorithm to analyze the user's emotional state. The input for this step is video data, and the output is numerical data representing emotions. The server analyzes the user's facial expressions and detects emotional patterns.

[0560] Step 5:

[0561] The server integrates natural language processing results and facial recognition results, and uses this to evaluate the potential danger of the spoken content. The input is text data and emotion data, and the output is a danger score as a harassment index. The server analyzes the fused data to score the potential harassment risk.

[0562] Step 6:

[0563] The server generates feedback information based on the risk score and sends it to the terminal. The input is the risk score, and the output is a feedback message to notify the user. The server then forms an appropriate advice or warning message.

[0564] Step 7:

[0565] The terminal notifies the user of feedback information received from the server via its display or speaker. The input is the feedback message, and the output is the transmission of information visually or audibly. This allows the user to immediately review their own actions.

[0566] Step 8:

[0567] The server stores all analysis results and feedback history in a database. The inputs are the analysis data and feedback history, and the output is the recorded data. The server stores this data for future retraining.

[0568] Step 9:

[0569] The server retrains a machine learning model based on stored data. The input is the stored historical data, and the output is the improved model. The server periodically updates the model to improve analysis accuracy and increase the reliability of the feedback.

[0570] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0571] This invention is a system that acquires user voice and video data, assesses the risk of harassment based on this data, and provides immediate feedback to the user. The system further incorporates an emotion engine to recognize the user's emotions.

[0572] First, the device uses its microphone and camera to capture data on the user's conversation and facial expressions. This data is securely processed using encryption technology and sent to the server. The transmitted data is guaranteed to be encrypted to ensure the user's privacy.

[0573] Next, the server converts the received audio data into text using natural language processing (NLP) technology. Based on this text information, it analyzes specific keywords and language patterns to assess the risk of harassment.

[0574] Furthermore, an emotion engine is used to analyze video data and identify the user's emotional state. This emotional information is integrated with linguistic information to enable more accurate risk assessment. This integrated system provides richer information than simple linguistic data analysis alone.

[0575] The server calculates a harassment score based on the analyzed data and determines the risk level (low, medium, or high). Based on this risk level, it generates a feedback message to notify the user. This feedback is sent to the device and provided to the user visually or audibly in real time.

[0576] For example, if a user uses aggressive language during a meeting, the device records the words and facial expressions, which are then analyzed by the server. If the emotion engine detects that the user is irritated, the server classifies it as high-risk and notifies the user via the device with feedback such as, "This statement is excessively aggressive."

[0577] The analysis data and feedback results are stored in a database. This allows the server to retrain its machine learning models based on the accumulated data, gradually improving the system's analysis accuracy. The emotion engine also continuously learns emotional patterns from past data, contributing to further improvements in the accuracy of harassment risk assessments in the future. This enables users to review their behavior more quickly and appropriately, preventing harassment proactively.

[0578] The following describes the processing flow.

[0579] Step 1:

[0580] When a user initiates a conversation, the device's microphone and camera activate, capturing audio and video data in real time. The captured data is temporarily stored within the device.

[0581] Step 2:

[0582] The device encrypts the acquired data, securely compresses it, and sends it to the server. The encrypted data is transferred through a privacy-conscious communication protocol.

[0583] Step 3:

[0584] The server converts the received audio data into text using a natural language processing (NLP) engine. This text data serves as the basis for evaluating harassment risk.

[0585] Step 4:

[0586] The server begins analyzing the text data and assesses the risk of the content of the statements by detecting specific keywords and phrase patterns.

[0587] Step 5:

[0588] The server uses video data to analyze the user's facial expressions with an emotion engine and recognize their emotional state (joy, anger, sadness, etc.). This emotion analysis allows for an understanding of the emotional context influencing their statements.

[0589] Step 6:

[0590] The server integrates the results of voice analysis and emotion analysis to calculate an overall harassment score. This score combines verbal content and emotional state, leading to a more accurate risk assessment.

[0591] Step 7:

[0592] The server classifies the risk level (low, medium, high) based on the obtained harassment score and generates a feedback message for the user. This message is adjusted to be appropriate according to the risk level.

[0593] Step 8:

[0594] A feedback message is sent from the server to the terminal, and the terminal notifies the user visually or audibly. The notification attracts the user's attention and prompts them to modify their behavior.

[0595] Step 9:

[0596] The server stores all analysis results and feedback history in a database, and uses this data to retrain the machine learning model, improving the system's analysis accuracy and the reliability of its sentiment recognition.

[0597] This processing flow allows users to receive immediate feedback on their actions, which can contribute to harassment prevention.

[0598] (Example 2)

[0599] Next, we will describe Example 2. 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."

[0600] This invention addresses the need to assess the risk of verbal and nonverbal harassment in real time in workplaces, educational institutions, and other settings, and to provide immediate feedback to users. Therefore, it requires technology that integrates audio and video analysis to capture subtle facial expressions and emotional changes that are often overlooked by conventional methods. However, current technologies have limitations in analysis accuracy and real-time capabilities, necessitating the development of systems with more advanced data analysis and feedback functions.

[0601] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0602] In this invention, the server includes means for integrating acoustic and image information to calculate a harassment score, determine a risk level, and generate a feedback message; means for notifying the user of the generated feedback message; and means for storing the analyzed information and feedback results. This enables real-time assessment of harassment risk and the provision of appropriate feedback to the user.

[0603] "Acoustic information" refers to sound data acquired using microphones and other sound acquisition devices, and includes human speech and ambient sounds.

[0604] "Image information" refers to visual data acquired using cameras and other video acquisition devices, and includes things like people's facial expressions and movements.

[0605] "Communication method" refers to protocols and means for securely sending and receiving data, including encryption technologies such as SSL / TLS.

[0606] A "natural language processing engine" is a technology that converts audio data into text, analyzes its content, and understands its meaning; it typically uses AI models.

[0607] A "facial expression recognition algorithm" is a technology that analyzes human facial expressions from image data and estimates their emotional state, utilizing deep learning models.

[0608] A "harassment score" is an index that quantitatively expresses the risk of harassment based on information obtained from audio and video.

[0609] A "feedback message" is a message containing instructions or comments that include points of criticism or warnings provided to the user based on the analyzed data.

[0610] A "computational model" is a mathematical structure that uses techniques such as machine learning to learn features from large amounts of data and perform predictions and classifications.

[0611] This invention is a system that uses acoustic and visual information to assess the risk of harassment and provides real-time feedback to the user. First, the terminal is equipped with a microphone and camera to capture the user's conversation and facial expressions. A high-sensitivity microphone and high-resolution camera are used for this purpose, enabling noise suppression and acquisition of high-definition images. The acquired data is encrypted using AES encryption technology and transmitted to the server using the SSL / TLS protocol.

[0612] The server performs a process of converting received acoustic information into text using a natural language processing engine. This process employs common speech recognition technologies. Next, an AI model is used to analyze the text data and assess risk based on specific keywords and language patterns.

[0613] Furthermore, the server analyzes the user's emotional state based on image information using facial recognition algorithms. This process utilizes libraries such as OpenCV and deep learning models to identify emotions from subtle facial expressions. The resulting emotional information and acoustic information are then integrated to assess risk in the form of a harassment score.

[0614] If a user uses offensive language during a meeting, the device records the user's facial expressions along with the words, and these are analyzed in detail on the server. If the emotion engine detects that the user is upset, the situation is assessed as high risk, and a feedback message such as "This statement is excessively offensive" is generated. This feedback is immediately communicated to the user visually or audibly through the device.

[0615] An example of a prompt message in this invention would be, "Please tell us about a situation in a workplace conversation that you found unpleasant. Please describe in detail what was said and how your feelings changed at that time." This allows users to objectively evaluate their own behavior and strive for more appropriate communication.

[0616] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0617] Step 1:

[0618] The device uses a microphone and camera to collect the user's voice and video. Audio data is acquired clearly using technology that effectively removes noise, and video data is captured in high resolution to record the user's facial expressions in detail. The input in this step is the user's raw voice and video, which are output as digital data.

[0619] Step 2:

[0620] The terminal encrypts the acquired digital data using AES encryption technology and sends it to the server using the secure protocol SSL / TLS. Security techniques are implemented during this process to ensure the safe transfer of highly confidential data. The input is the digital data before encryption, and the output is the encrypted digital data.

[0621] Step 3:

[0622] The server converts the received audio data into text data using a natural language processing engine. Specifically, it uses speech recognition technology to convert the audio signal into text information. The input is encrypted audio data, and the output is parseable text data.

[0623] Step 4:

[0624] The server performs a risk assessment using an AI model based on the converted text data. In this step, keywords and contextual patterns are analyzed to quantify the risk of harassment. This analysis process generates a risk score as output from the text data as input.

[0625] Step 5:

[0626] The server applies a facial recognition algorithm to video data to analyze the user's emotional state. Using a deep learning model, it infers emotions from subtle changes in facial expressions. The input is encrypted video data, and the output is an emotion score.

[0627] Step 6:

[0628] The server integrates data obtained from audio and video to calculate a harassment score. This integrated analysis enables more accurate harassment assessment than conventional methods. The input consists of a risk score from audio and an emotion score from images, and the output is the integrated harassment score.

[0629] Step 7:

[0630] The server assesses the risk level based on the obtained harassment score and generates a feedback message. For example, in the case of a high-risk situation, it provides appropriate feedback such as "This statement is excessively aggressive." The input is the harassment score, and the output is the generated feedback message.

[0631] Step 8:

[0632] The terminal notifies the user of the generated feedback message. Warnings and suggestions are communicated immediately using a visual display or audio output device. The input is the feedback message sent from the server, and the output is the notification to the user.

[0633] Step 9:

[0634] The server records the analyzed data and feedback history in a database. This enables long-term data analysis and contributes to improving the system's accuracy. The input is the analysis results and feedback history, and the output is the updated database.

[0635] Step 10:

[0636] The server retrains the generative AI model based on the accumulated data. This process refines the machine learning algorithm, enabling it to continue learning from new data patterns. The input is stored historical data, and the output is the improved AI model.

[0637] (Application Example 2)

[0638] Next, we will explain Application Example 2. In the following explanation, 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."

[0639] In brick-and-mortar stores, there is a lack of systems to immediately detect the risk of harassment in communication between staff and customers and to enable staff to take appropriate action. This problem can lead to misunderstandings and inappropriate responses between customers, potentially resulting in a decline in service quality and unnecessary trouble. Therefore, a system that ensures safe and smooth communication is needed.

[0640] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0641] In this invention, the server includes an acquisition means for acquiring audio and video information; a transmission means for transmitting the acquired audio and video information via a communication standard; an evaluation means for analyzing the received audio information with a natural language processing device and evaluating the degree of danger of the spoken content; an analysis means for analyzing emotional states from video information using facial expression recognition calculations; a generation means for determining the degree of danger based on harassment indicators and generating feedback information; a notification means for notifying the user of the generated feedback information; a storage means for storing analysis information and history; a training means for retraining a machine learning model based on the stored information; and a notification means using a visual device to detect signs of harassment in real time and provide feedback to staff. This makes it possible to immediately assess the risk of harassment during conversations in physical stores and send feedback to staff.

[0642] "Audio and video information" refers to digital data that records the user's speech, facial expressions, and actions.

[0643] "Means of acquisition" refers to devices and technologies for collecting audio and video information.

[0644] "Communication standards" refer to protocols and specifications for sending and receiving data.

[0645] "Transmission means" refers to devices or technologies used to send acquired information to other devices such as servers.

[0646] A "natural language processing system" is a technology that analyzes speech information and understands it as text data.

[0647] "Evaluation methods" refer to devices and technologies used to analyze acquired information and determine its degree of risk.

[0648] "Facial expression recognition calculation" is an algorithm used to analyze an individual's emotions and state of mind from video information.

[0649] "Analysis means" refers to devices and technologies used to analyze acquired information in detail and extract its meaning.

[0650] "Harassment indicators" are criteria or indicators used to evaluate whether communication constitutes harassment.

[0651] "Generation means" refers to devices or technologies for forming feedback information based on analysis results.

[0652] "Notification means" refers to devices or technologies used to transmit generated feedback information to the user.

[0653] "Storage methods" refer to devices and technologies for storing analysis results and historical data.

[0654] "Training methods" refer to techniques for updating machine learning models using stored data.

[0655] A "visual device" is a device that allows users to receive feedback information visually.

[0656] The system realizing this invention acquires audio and video information in real time and evaluates the risk of harassment based on that information. Within the target physical store, staff wear smart glasses and communicate with a server via wireless communication to exchange information. The smart glasses have a built-in camera and microphone, which are used to acquire audio and video information.

[0657] The server converts audio data into text using a natural language processing unit (e.g., Google Cloud Speech-to-Text) and analyzes the content of the speech. Simultaneously, video data is analyzed by facial expression recognition software (e.g., Microsoft Azure Emotion API) to assess the emotional state of customers and staff. These analysis results are comprehensively evaluated based on harassment indicators to determine the level of risk.

[0658] Based on the risk level, the server immediately generates feedback information and displays it on the smart glasses' display to notify staff. This visual notification provides staff with guidance for taking appropriate action. Furthermore, the analyzed information is stored in a secure database and used to retrain machine learning models. Through this process, the system improves in accuracy over time, enabling it to provide more precise feedback.

[0659] For example, if a store detects signs of customer dissatisfaction, staff can receive feedback through smart glasses such as, "The customer's tone is agitated. Please try to remain calm." This allows staff to calmly assess the situation and improve service quality.

[0660] An example of a prompt message is: "If recent customer interactions have detected voice and facial expressions indicating increased stress, what measures should be taken? What message should be displayed in the text notification?"

[0661] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0662] Step 1:

[0663] The device (smart glasses) acquires real-time audio and video information of staff and customers within the store. The input is raw data from the camera and microphone, which is then formatted as digital information. Audio files and video clips are generated as output.

[0664] Step 2:

[0665] The terminal transmits the acquired audio and video information to the server. The input is the digital information generated in step 1. The data is encrypted during transmission and securely sent to the server via communication standards. The server receives the encrypted information as output.

[0666] Step 3:

[0667] The server analyzes the received audio data using a natural language processing unit. The input is an encrypted audio file, which is decrypted and converted into text data. The output is the transcribed speech content, which is then analyzed to prepare for risk assessment.

[0668] Step 4:

[0669] The video data is analyzed on a server using facial recognition software. The input is an encrypted video clip, which is decrypted to evaluate the emotional state of staff and customers. The output is the analyzed emotional data, which is used for risk assessment.

[0670] Step 5:

[0671] The server integrates transcribed speech content and sentiment data, and evaluates the risk level based on harassment indicators. The input is the analysis results from steps 3 and 4. The data is integrated to determine the risk level (low, medium, high). The output is the risk level and the feedback information that should be taken.

[0672] Step 6:

[0673] The server generates feedback information and notifies the terminal. The input is the feedback information determined in step 5. This is sent to the terminal's display to provide visual feedback to the staff. The output is a feedback display on the terminal's display that is recognizable to the staff.

[0674] Step 7:

[0675] The server encrypts the analysis data and risk assessment history and stores it in a database. The input is the assessment results and feedback history from step 5. This is saved and used as training data for future machine learning models. The output is the updated database.

[0676] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0677] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0678] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0679] [Fourth Embodiment]

[0680] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0681] As shown in Figure 7, the 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.

[0682] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0683] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0684] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0685] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0686] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0687] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive 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 robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0688] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0689] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0690] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0691] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0692] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0693] The system of the present invention acquires audio and video data in real time, analyzes it to assess the risk of harassment, and provides feedback to the user.

[0694] In this embodiment, the terminal first uses a microphone and camera to acquire the user's voice and video as data. This data is transmitted to the server in an encrypted format. To protect privacy, advanced encryption protocols are used for communication during data transmission.

[0695] The server analyzes the received data. First, it converts audio data into text using natural language processing (NLP) technology, and then measures the likelihood of harassment based on the text structure and content. Simultaneously, it analyzes the user's emotional state using facial recognition algorithms based on video data. This analysis allows for a multifaceted evaluation of the user's intentions and how they were perceived.

[0696] Next, the server scores the harassment risk based on the analysis results and determines the risk level (low, medium, or high) based on that score. It then generates an appropriate feedback message accordingly. This generated feedback is returned to the terminal in real time and notified to the user. Based on the feedback, the user can immediately review their own words and actions.

[0697] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and facial expressions and sends them to the server. The server assigns a high risk score to this comment, citing gender bias, and notifies the user through the device that "this comment may be inappropriate."

[0698] Furthermore, the server stores a history of all analysis and feedback, and uses this to retrain the machine learning model, improving accuracy and reliability. This process allows the system to continuously provide intuitive and useful feedback to users, thereby preventing harassment and improving the quality of communication.

[0699] The following describes the processing flow.

[0700] Step 1:

[0701] The device uses a microphone and camera to acquire the user's voice and video data. The data is captured in real time and temporarily stored by the system.

[0702] Step 2:

[0703] The device compresses the acquired audio and video data and sends it to the server in a highly encrypted format. This enhances data security.

[0704] Step 3:

[0705] The server uses a natural language processing (NLP) engine to convert the received audio data into text. Next, it identifies specific keywords and language patterns from the analyzed text to assess the potential for harassment.

[0706] Step 4:

[0707] The server analyzes the received video data using a facial recognition algorithm to evaluate the user's emotional state. This information is combined with text analysis results to perform a comprehensive risk assessment.

[0708] Step 5:

[0709] The server analyzes audio and video data to score the harassment risk and classifies it into three risk levels (low, medium, high). It then generates a feedback message corresponding to this risk level.

[0710] Step 6:

[0711] The server sends the generated feedback message to the terminal.

[0712] Step 7:

[0713] The device notifies the user of the generated feedback message in real time. This gives the user an opportunity to review their own words and actions.

[0714] Step 8:

[0715] The server stores the analysis results and user feedback history in a database.

[0716] Step 9:

[0717] The server uses the stored data to retrain the machine learning model and improve the system's accuracy. This process is performed periodically.

[0718] (Example 1)

[0719] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0720] In today's communication environment, it is essential to recognize the risks of harassment early and for users to appropriately avoid them. However, conventional technologies do not adequately provide efficient methods for detecting harassment using audio and video, and there is a lack of mechanisms to provide real-time feedback on inappropriate remarks and actions that users may unconsciously make. As a result, the detection and correction of harassment is delayed, hindering the improvement of the quality of communication.

[0721] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0722] In this invention, the server includes acquisition means for acquiring acoustic and image data, transmission means for transmitting this data via an information transmission protocol, and analysis means that use a natural language processing engine and image recognition algorithms to analyze the received data, assess the risks, and provide feedback. This enables early detection of harassment risks, and allows users to receive real-time feedback, enabling them to consciously and quickly review their behavior.

[0723] "Audio data" refers to digital or analog data containing audio information, acquired for the purpose of transmitting information based on sound waves.

[0724] "Image data" refers to digital or analog data containing visual information acquired by cameras or other imaging devices.

[0725] An "information transmission protocol" is a set of rules and procedures for communicating data securely and efficiently.

[0726] A "natural language processing engine" is a set of algorithms that analyze speech or text data to support the understanding and generation of natural language used by humans.

[0727] An "image recognition algorithm" is an algorithm that analyzes specific patterns or features within an image to extract and recognize information.

[0728] A "harassment evaluation index" is a standard or criterion used to quantify the possibility and degree of harassment.

[0729] A "computer learning model" is a framework of algorithms that learns patterns from data and performs predictions and classifications.

[0730] In an embodiment of the present invention, a terminal first acquires acoustic and image data using a microphone and a camera. This includes commonly used electronic devices, voice recognition microphones, and HD cameras. This information is encrypted in real time and transmitted to a server via an information transmission protocol.

[0731] Next, the server converts the received audio data into text data using a natural language processing engine. This engine includes, for example, speech recognition software. The server then analyzes the text data and assesses the potential risks based on the content of the speech. Simultaneously, the server analyzes the image data using an image recognition algorithm to determine the user's emotional state. This image recognition algorithm utilizes facial recognition software.

[0732] Based on these analysis results, the server generates harassment evaluation metrics and calculates a risk score in real time. If a high risk is determined, the AI ​​model generates an appropriate feedback message for the user and notifies them through their device.

[0733] For example, if a user makes a comment during a meeting such as, "This job is impossible for women anyway," the device records the audio and video information and sends it to the server. The server determines that the comment contains gender bias and scores it as high risk. It then sends feedback to the device stating, "This comment may be inappropriate," and notifies the user. This notification allows the user to immediately review their comment.

[0734] An example of a prompt message would be: "Voice: This job is probably impossible for a woman. Emotional state: Neutral. Please analyze."

[0735] This system allows users to recognize the risk of harassment they may have unintentionally caused early on and have the opportunity to improve it. Furthermore, the server stores all analysis results and feedback history, and uses this data to retrain the computer learning model, continuously improving the overall accuracy and reliability of the system.

[0736] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0737] Step 1:

[0738] The device acquires audio and image data from the user. It records audio with a microphone and captures images with a camera. This acquired data serves as input. The data is encrypted and sent to the server using an information transmission protocol. The output consists of encrypted audio and image data.

[0739] Step 2:

[0740] The server inputs the received audio data into a natural language processing engine, which converts it into text data. Here, speech recognition technology is used to analyze the language from the audio waveform and transcribe it into text. This process outputs the audio content in text format.

[0741] Step 3:

[0742] Simultaneously, the server uses an image recognition algorithm to input image data and analyze the user's emotional state. During the analysis process, it captures feature points of facial expressions, identifies subtle changes in facial expression, and outputs the emotional state (e.g., joy, anger, sadness).

[0743] Step 4:

[0744] The server integrates text data and image sentiment analysis results, and uses this information to calculate a harassment assessment index. Specifically, it utilizes a generative AI model to input text and sentiment information and score the risk level of the remarks. The output is a numerical score of harassment risk.

[0745] Step 5:

[0746] The server generates feedback information based on the calculated risk score. Using a generation AI model, it creates text messages corresponding to the risk level, and the content of those messages is output.

[0747] Step 6:

[0748] The server sends the generated feedback information to the terminal. The terminal notifies the user of the received feedback, allowing the user to modify their actions based on it. Real-time notifications to the user are output.

[0749] Step 7:

[0750] The server stores all analysis results and feedback history. This dataset is used to retrain the computer learning model, continuously improving the system's accuracy and reliability. Data storage and model training are the outputs.

[0751] (Application Example 1)

[0752] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0753] In family communication, unintentional harassment and inappropriate remarks can damage relationships with family members and guests. Preventing such problems and maintaining a smooth and harmonious communication environment is crucial. However, currently, there is a lack of means to detect these problems in real time and provide feedback. Therefore, there is a need for a system that can address this challenge.

[0754] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0755] In this invention, the server includes acquisition means for acquiring acoustic and visual information; transmission means for transmitting the acquired acoustic and visual information via a communication protocol; evaluation means for analyzing the received acoustic information with a natural language processing engine and evaluating the risk of the content of the speech; analysis means for analyzing emotional states from visual information using an facial expression recognition algorithm; generation means for determining the degree of risk based on harassment indicators and generating feedback information; notification means for notifying the user of the generated feedback information; storage means for storing analysis information and history information; training means for retraining a machine learning model based on the stored information; and means having a function to monitor conversations within the home, detect specific phrases to support smooth communication, and provide feedback. This makes it possible to improve the quality of communication within the home.

[0756] "Acoustic information" refers to data used to acquire ambient sounds and speech.

[0757] "Visual information" refers to video data acquired through a camera.

[0758] A "communication protocol" is a set of rules that define how data can be transmitted securely.

[0759] A "natural language processing engine" is a technology that converts audio data into text and analyzes its content.

[0760] A "facial expression recognition algorithm" is a method for analyzing a user's emotional state from video data.

[0761] A "harassment index" is a set of values ​​used as a standard for evaluating the risks of words and actions.

[0762] "Feedback information" refers to advice and warning messages provided to users based on the analysis results.

[0763] A "machine learning model" is a collection of algorithms used to perform analysis and predictions based on data.

[0764] "Monitoring conversations within the household" means monitoring the speech and actions of members of the household to detect potentially dangerous behaviors.

[0765] This invention is a system designed to improve the quality of communication within the home. To acquire acoustic and visual information, the server uses a terminal equipped with a high-performance microphone and camera. This allows for the acquisition of acoustic and video data in real time. This data is transmitted to the server via a communication protocol. The communication protocol uses encryption technologies such as TLS / SSL to ensure secure data transmission.

[0766] On the server, received audio data is converted into text using a natural language processing engine (e.g., the spaCy library) and the potential danger of the spoken content is evaluated. For visual information, the user's emotional state is analyzed using an facial recognition algorithm based on OpenCV. Based on these analysis results, a harassment index is calculated and the degree of danger is evaluated.

[0767] Based on the assessed risk level, the server generates feedback information. This generated feedback information is then notified to the user. Since the feedback information is presented in real time through the device's display or speaker, the user can immediately review their own statements and actions.

[0768] Furthermore, the server stores analysis information and feedback history. Based on this stored information, the machine learning model can be retrained to improve the accuracy and reliability of the feedback. By detecting specific phrases in conversations within the home and providing appropriate feedback, it supports smoother communication.

[0769] For example, if someone says "My opinion might be considered boring" during a family dinner, the server can pick up on this phrase, analyze the emotion behind it, and then provide feedback such as, "Why not try trusting yourself more and sharing your opinion?"

[0770] An example of a prompt to input into a generative AI model is, "Please describe how to assess the harassment risk in conversations within the family and provide feedback on specific areas for improvement."

[0771] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0772] Step 1:

[0773] The device uses a high-performance microphone and camera to acquire acoustic and visual information within the home. This input consists of audio and video data collected in real time. The device stores this data in a buffer and prepares it for subsequent processing.

[0774] Step 2:

[0775] The terminal transmits the acquired audio and visual information to the server via a secure communication protocol using encryption technologies such as TLS / SSL. The input data is raw audio and video, and the output is this data encrypted and sent to the server. The server decrypts the received data and makes it analyzable.

[0776] Step 3:

[0777] The server converts the received audio data into text using a natural language processing engine (e.g., spaCy). The input for this step is audio data, and the output is the generated text data. The server then performs speech recognition and processes the conversation content as textual information.

[0778] Step 4:

[0779] The server processes the received visual information using an OpenCV-based facial recognition algorithm to analyze the user's emotional state. The input for this step is video data, and the output is numerical data representing emotions. The server analyzes the user's facial expressions and detects emotional patterns.

[0780] Step 5:

[0781] The server integrates natural language processing results and facial recognition results, and uses this to evaluate the potential danger of the spoken content. The input is text data and emotion data, and the output is a danger score as a harassment index. The server analyzes the fused data to score the potential harassment risk.

[0782] Step 6:

[0783] The server generates feedback information based on the risk score and sends it to the terminal. The input is the risk score, and the output is a feedback message to notify the user. The server then forms an appropriate advice or warning message.

[0784] Step 7:

[0785] The terminal notifies the user of feedback information received from the server via its display or speaker. The input is the feedback message, and the output is the transmission of information visually or audibly. This allows the user to immediately review their own actions.

[0786] Step 8:

[0787] The server stores all analysis results and feedback history in a database. The inputs are the analysis data and feedback history, and the output is the recorded data. The server stores this data for future retraining.

[0788] Step 9:

[0789] The server retrains a machine learning model based on stored data. The input is the stored historical data, and the output is the improved model. The server periodically updates the model to improve analysis accuracy and increase the reliability of the feedback.

[0790] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0791] This invention is a system that acquires user voice and video data, assesses the risk of harassment based on this data, and provides immediate feedback to the user. The system further incorporates an emotion engine to recognize the user's emotions.

[0792] First, the device uses its microphone and camera to capture data on the user's conversation and facial expressions. This data is securely processed using encryption technology and sent to the server. The transmitted data is guaranteed to be encrypted to ensure the user's privacy.

[0793] Next, the server converts the received audio data into text using natural language processing (NLP) technology. Based on this text information, it analyzes specific keywords and language patterns to assess the risk of harassment.

[0794] Furthermore, an emotion engine is used to analyze video data and identify the user's emotional state. This emotional information is integrated with linguistic information to enable more accurate risk assessment. This integrated system provides richer information than simple linguistic data analysis alone.

[0795] The server calculates a harassment score based on the analyzed data and determines the risk level (low, medium, or high). Based on this risk level, it generates a feedback message to notify the user. This feedback is sent to the device and provided to the user visually or audibly in real time.

[0796] For example, if a user uses aggressive language during a meeting, the device records the words and facial expressions, which are then analyzed by the server. If the emotion engine detects that the user is irritated, the server classifies it as high-risk and notifies the user via the device with feedback such as, "This statement is excessively aggressive."

[0797] The analysis data and feedback results are stored in a database. This allows the server to retrain its machine learning models based on the accumulated data, gradually improving the system's analysis accuracy. The emotion engine also continuously learns emotional patterns from past data, contributing to further improvements in the accuracy of harassment risk assessments in the future. This enables users to review their behavior more quickly and appropriately, preventing harassment proactively.

[0798] The following describes the processing flow.

[0799] Step 1:

[0800] When a user initiates a conversation, the device's microphone and camera activate, capturing audio and video data in real time. The captured data is temporarily stored within the device.

[0801] Step 2:

[0802] The device encrypts the acquired data, securely compresses it, and sends it to the server. The encrypted data is transferred through a privacy-conscious communication protocol.

[0803] Step 3:

[0804] The server converts the received audio data into text using a natural language processing (NLP) engine. This text data serves as the basis for evaluating harassment risk.

[0805] Step 4:

[0806] The server begins analyzing the text data and assesses the risk of the content of the statements by detecting specific keywords and phrase patterns.

[0807] Step 5:

[0808] The server uses video data to analyze the user's facial expressions with an emotion engine and recognize their emotional state (joy, anger, sadness, etc.). This emotion analysis allows for an understanding of the emotional context influencing their statements.

[0809] Step 6:

[0810] The server integrates the results of voice analysis and emotion analysis to calculate an overall harassment score. This score combines verbal content and emotional state, leading to a more accurate risk assessment.

[0811] Step 7:

[0812] The server classifies the risk level (low, medium, high) based on the obtained harassment score and generates a feedback message for the user. This message is adjusted to be appropriate according to the risk level.

[0813] Step 8:

[0814] A feedback message is sent from the server to the terminal, and the terminal notifies the user visually or audibly. The notification attracts the user's attention and prompts them to modify their behavior.

[0815] Step 9:

[0816] The server stores all analysis results and feedback history in a database, and uses this data to retrain the machine learning model, improving the system's analysis accuracy and the reliability of its sentiment recognition.

[0817] This processing flow allows users to receive immediate feedback on their actions, which can contribute to harassment prevention.

[0818] (Example 2)

[0819] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0820] This invention addresses the need to assess the risk of verbal and nonverbal harassment in real time in workplaces, educational institutions, and other settings, and to provide immediate feedback to users. Therefore, it requires technology that integrates audio and video analysis to capture subtle facial expressions and emotional changes that are often overlooked by conventional methods. However, current technologies have limitations in analysis accuracy and real-time capabilities, necessitating the development of systems with more advanced data analysis and feedback functions.

[0821] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0822] In this invention, the server includes means for integrating acoustic and image information to calculate a harassment score, determine a risk level, and generate a feedback message; means for notifying the user of the generated feedback message; and means for storing the analyzed information and feedback results. This enables real-time assessment of harassment risk and the provision of appropriate feedback to the user.

[0823] "Acoustic information" refers to sound data acquired using microphones and other sound acquisition devices, and includes human speech and ambient sounds.

[0824] "Image information" refers to visual data acquired using cameras and other video acquisition devices, and includes things like people's facial expressions and movements.

[0825] "Communication method" refers to protocols and means for securely sending and receiving data, including encryption technologies such as SSL / TLS.

[0826] A "natural language processing engine" is a technology that converts audio data into text, analyzes its content, and understands its meaning; it typically uses AI models.

[0827] A "facial expression recognition algorithm" is a technology that analyzes human facial expressions from image data and estimates their emotional state, utilizing deep learning models.

[0828] A "harassment score" is an index that quantitatively expresses the risk of harassment based on information obtained from audio and video.

[0829] A "feedback message" is a message containing instructions or comments that include points of criticism or warnings provided to the user based on the analyzed data.

[0830] A "computational model" is a mathematical structure that uses techniques such as machine learning to learn features from large amounts of data and perform predictions and classifications.

[0831] This invention is a system that uses acoustic and visual information to assess the risk of harassment and provides real-time feedback to the user. First, the terminal is equipped with a microphone and camera to capture the user's conversation and facial expressions. A high-sensitivity microphone and high-resolution camera are used for this purpose, enabling noise suppression and acquisition of high-definition images. The acquired data is encrypted using AES encryption technology and transmitted to the server using the SSL / TLS protocol.

[0832] The server performs a process of converting received acoustic information into text using a natural language processing engine. This process employs common speech recognition technologies. Next, an AI model is used to analyze the text data and assess risk based on specific keywords and language patterns.

[0833] Furthermore, the server analyzes the user's emotional state based on image information using facial recognition algorithms. This process utilizes libraries such as OpenCV and deep learning models to identify emotions from subtle facial expressions. The resulting emotional information and acoustic information are then integrated to assess risk in the form of a harassment score.

[0834] If a user uses offensive language during a meeting, the device records the user's facial expressions along with the words, and these are analyzed in detail on the server. If the emotion engine detects that the user is upset, the situation is assessed as high risk, and a feedback message such as "This statement is excessively offensive" is generated. This feedback is immediately communicated to the user visually or audibly through the device.

[0835] An example of a prompt message in this invention would be, "Please tell us about a situation in a workplace conversation that you found unpleasant. Please describe in detail what was said and how your feelings changed at that time." This allows users to objectively evaluate their own behavior and strive for more appropriate communication.

[0836] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0837] Step 1:

[0838] The device uses a microphone and camera to collect the user's voice and video. Audio data is acquired clearly using technology that effectively removes noise, and video data is captured in high resolution to record the user's facial expressions in detail. The input in this step is the user's raw voice and video, which are output as digital data.

[0839] Step 2:

[0840] The terminal encrypts the acquired digital data using AES encryption technology and sends it to the server using the secure protocol SSL / TLS. Security techniques are implemented during this process to ensure the safe transfer of highly confidential data. The input is the digital data before encryption, and the output is the encrypted digital data.

[0841] Step 3:

[0842] The server converts the received audio data into text data using a natural language processing engine. Specifically, it uses speech recognition technology to convert the audio signal into text information. The input is encrypted audio data, and the output is parseable text data.

[0843] Step 4:

[0844] The server performs a risk assessment using an AI model based on the converted text data. In this step, keywords and contextual patterns are analyzed to quantify the risk of harassment. This analysis process generates a risk score as output from the text data as input.

[0845] Step 5:

[0846] The server applies a facial recognition algorithm to video data to analyze the user's emotional state. Using a deep learning model, it infers emotions from subtle changes in facial expressions. The input is encrypted video data, and the output is an emotion score.

[0847] Step 6:

[0848] The server integrates data obtained from audio and video to calculate a harassment score. This integrated analysis enables more accurate harassment assessment than conventional methods. The input consists of a risk score from audio and an emotion score from images, and the output is the integrated harassment score.

[0849] Step 7:

[0850] The server assesses the risk level based on the obtained harassment score and generates a feedback message. For example, in the case of a high-risk situation, it provides appropriate feedback such as "This statement is excessively aggressive." The input is the harassment score, and the output is the generated feedback message.

[0851] Step 8:

[0852] The terminal notifies the user of the generated feedback message. Warnings and suggestions are communicated immediately using a visual display or audio output device. The input is the feedback message sent from the server, and the output is the notification to the user.

[0853] Step 9:

[0854] The server records the analyzed data and feedback history in a database. This enables long-term data analysis and contributes to improving the system's accuracy. The input is the analysis results and feedback history, and the output is the updated database.

[0855] Step 10:

[0856] The server retrains the generative AI model based on the accumulated data. This process refines the machine learning algorithm, enabling it to continue learning from new data patterns. The input is stored historical data, and the output is the improved AI model.

[0857] (Application Example 2)

[0858] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0859] In brick-and-mortar stores, there is a lack of systems to immediately detect the risk of harassment in communication between staff and customers and to enable staff to take appropriate action. This problem can lead to misunderstandings and inappropriate responses between customers, potentially resulting in a decline in service quality and unnecessary trouble. Therefore, a system that ensures safe and smooth communication is needed.

[0860] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0861] In this invention, the server includes an acquisition means for acquiring audio and video information; a transmission means for transmitting the acquired audio and video information via a communication standard; an evaluation means for analyzing the received audio information with a natural language processing device and evaluating the degree of danger of the spoken content; an analysis means for analyzing emotional states from video information using facial expression recognition calculations; a generation means for determining the degree of danger based on harassment indicators and generating feedback information; a notification means for notifying the user of the generated feedback information; a storage means for storing analysis information and history; a training means for retraining a machine learning model based on the stored information; and a notification means using a visual device to detect signs of harassment in real time and provide feedback to staff. This makes it possible to immediately assess the risk of harassment during conversations in physical stores and send feedback to staff.

[0862] "Audio and video information" refers to digital data that records the user's speech, facial expressions, and actions.

[0863] "Means of acquisition" refers to devices and technologies for collecting audio and video information.

[0864] "Communication standards" refer to protocols and specifications for sending and receiving data.

[0865] "Transmission means" refers to devices or technologies used to send acquired information to other devices such as servers.

[0866] A "natural language processing system" is a technology that analyzes speech information and understands it as text data.

[0867] "Evaluation methods" refer to devices and technologies used to analyze acquired information and determine its degree of risk.

[0868] "Facial expression recognition calculation" is an algorithm used to analyze an individual's emotions and state of mind from video information.

[0869] "Analysis means" refers to devices and technologies used to analyze acquired information in detail and extract its meaning.

[0870] "Harassment indicators" are criteria or indicators used to evaluate whether communication constitutes harassment.

[0871] "Generation means" refers to devices or technologies for forming feedback information based on analysis results.

[0872] "Notification means" refers to devices or technologies used to transmit generated feedback information to the user.

[0873] "Storage methods" refer to devices and technologies for storing analysis results and historical data.

[0874] "Training methods" refer to techniques for updating machine learning models using stored data.

[0875] A "visual device" is a device that allows users to receive feedback information visually.

[0876] The system realizing this invention acquires audio and video information in real time and evaluates the risk of harassment based on that information. Within the target physical store, staff wear smart glasses and communicate with a server via wireless communication to exchange information. The smart glasses have a built-in camera and microphone, which are used to acquire audio and video information.

[0877] The server converts audio data into text using a natural language processing unit (e.g., Google Cloud Speech-to-Text) and analyzes the content of the speech. Simultaneously, video data is analyzed by facial expression recognition software (e.g., Microsoft Azure Emotion API) to assess the emotional state of customers and staff. These analysis results are comprehensively evaluated based on harassment indicators to determine the level of risk.

[0878] Based on the risk level, the server immediately generates feedback information and displays it on the smart glasses' display to notify staff. This visual notification provides staff with guidance for taking appropriate action. Furthermore, the analyzed information is stored in a secure database and used to retrain machine learning models. Through this process, the system improves in accuracy over time, enabling it to provide more precise feedback.

[0879] For example, if a store detects signs of customer dissatisfaction, staff can receive feedback through smart glasses such as, "The customer's tone is agitated. Please try to remain calm." This allows staff to calmly assess the situation and improve service quality.

[0880] An example of a prompt message is: "If recent customer interactions have detected voice and facial expressions indicating increased stress, what measures should be taken? What message should be displayed in the text notification?"

[0881] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0882] Step 1:

[0883] The device (smart glasses) acquires real-time audio and video information of staff and customers within the store. The input is raw data from the camera and microphone, which is then formatted as digital information. Audio files and video clips are generated as output.

[0884] Step 2:

[0885] The terminal transmits the acquired audio and video information to the server. The input is the digital information generated in step 1. The data is encrypted during transmission and securely sent to the server via communication standards. The server receives the encrypted information as output.

[0886] Step 3:

[0887] The server analyzes the received audio data using a natural language processing unit. The input is an encrypted audio file, which is decrypted and converted into text data. The output is the transcribed speech content, which is then analyzed to prepare for risk assessment.

[0888] Step 4:

[0889] The video data is analyzed on a server using facial recognition software. The input is an encrypted video clip, which is decrypted to evaluate the emotional state of staff and customers. The output is the analyzed emotional data, which is used for risk assessment.

[0890] Step 5:

[0891] The server integrates transcribed speech content and sentiment data, and evaluates the risk level based on harassment indicators. The input is the analysis results from steps 3 and 4. The data is integrated to determine the risk level (low, medium, high). The output is the risk level and the feedback information that should be taken.

[0892] Step 6:

[0893] The server generates feedback information and notifies the terminal. The input is the feedback information determined in step 5. This is sent to the terminal's display to provide visual feedback to the staff. The output is a feedback display on the terminal's display that is recognizable to the staff.

[0894] Step 7:

[0895] The server encrypts the analysis data and risk assessment history and stores it in a database. The input is the assessment results and feedback history from step 5. This is saved and used as training data for future machine learning models. The output is the updated database.

[0896] 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 controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0897] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0898] 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 this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0899] Furthermore, the emotion identification model 59, acting 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 a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0900] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0901] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0902] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0903] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0904] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is 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 the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0905] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0906] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0907] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0908] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0909] 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.

[0910] Furthermore, it is not necessary to store the entirety 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 the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0911] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0912] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of 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). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0913] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0914] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0915] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0916] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[0917] The following is further disclosed regarding the embodiments described above.

[0918] (Claim 1)

[0919] means for acquiring audio and video data,

[0920] A transmission means for transmitting acquired audio and video data via a communication protocol,

[0921] An evaluation method that analyzes received audio data using a natural language processing engine and evaluates the risk of the content of the speech,

[0922] An analysis method that analyzes emotional states from video data using a facial recognition algorithm,

[0923] A generation means that determines the risk level based on the harassment score and generates a feedback message,

[0924] A notification means for notifying the user of the generated feedback message,

[0925] A means of saving analysis data and history,

[0926] A training method for retraining a machine learning model based on stored data,

[0927] A system that includes this.

[0928] (Claim 2)

[0929] The system according to claim 1, wherein the transmitted audio and video data are encrypted before transmission.

[0930] (Claim 3)

[0931] The system according to claim 1, wherein the accuracy of the analyzed harassment risk is improved by comparing it with past data.

[0932] "Example 1"

[0933] (Claim 1)

[0934] Acquisition means for acquiring acoustic data and image data,

[0935] A transmission means for transmitting acquired acoustic data and image data via an information transmission protocol,

[0936] An evaluation method that analyzes received acoustic data using a natural language processing engine to assess the danger of the spoken content,

[0937] An analysis method that uses an image recognition algorithm to analyze emotional states from image data,

[0938] A generation means that determines the degree of risk based on harassment evaluation indicators and generates feedback information,

[0939] A notification means for notifying the user of the generated feedback information,

[0940] A storage means for storing analysis data and history,

[0941] A learning method for retraining a computer learning model based on stored data,

[0942] A system that includes this.

[0943] (Claim 2)

[0944] The system according to claim 1, wherein the transmitted audio data and image data are transmitted in an encrypted state.

[0945] (Claim 3)

[0946] The system according to claim 1, wherein the accuracy of the analyzed harassment risk is improved by comparing it with past data.

[0947] "Application Example 1"

[0948] (Claim 1)

[0949] Acquisition means for acquiring acoustic and visual information,

[0950] A transmission means for transmitting acquired acoustic and visual information via a communication protocol,

[0951] An evaluation method that analyzes received acoustic information using a natural language processing engine and evaluates the danger of the content of the speech,

[0952] An analytical means for analyzing emotional states from visual information using an facial expression recognition algorithm,

[0953] A generation means that determines the degree of risk based on harassment indicators and generates feedback information,

[0954] A notification means for notifying the user of the generated feedback information,

[0955] A storage means for storing analysis information and historical information,

[0956] A training method for retraining a machine learning model based on stored information,

[0957] A means having the function of monitoring conversations within the home, detecting specific phrases to support smooth communication, and providing feedback,

[0958] A system that includes this.

[0959] (Claim 2)

[0960] The system according to claim 1, wherein the transmitted acoustic and visual information is encrypted before transmission.

[0961] (Claim 3)

[0962] The system according to claim 1, wherein the accuracy of the analyzed harassment risk is improved by comparing it with past information.

[0963] "Example 2 of combining an emotion engine"

[0964] (Claim 1)

[0965] Acquisition means for acquiring acoustic information and image information,

[0966] A transmission means for transmitting acquired acoustic and image information via a communication method,

[0967] An evaluation method that analyzes received acoustic information using a natural language processing engine and evaluates the risk of the content of the speech,

[0968] An analysis method that analyzes emotional states from image information using an facial expression recognition algorithm,

[0969] A generation means that integrates acoustic and emotional information to calculate a harassment score, determines the risk level, and generates a feedback message,

[0970] A notification means for notifying the user of the generated feedback message,

[0971] A storage means for storing the analyzed information and feedback results,

[0972] A training method that retrains the computational model based on memorized information to sequentially improve the accuracy of the analysis,

[0973] A system that includes this.

[0974] (Claim 2)

[0975] The system according to claim 1, wherein the transmitted acoustic information and image information are transmitted in an encrypted state.

[0976] (Claim 3)

[0977] The system according to claim 1, wherein the accuracy of the analyzed harassment risk is improved by comparing it with past information.

[0978] "Application example 2 when combining with an emotional engine"

[0979] (Claim 1)

[0980] means for acquiring audio and video information,

[0981] A transmission means for transmitting acquired audio and video information via a communication standard,

[0982] An evaluation means that analyzes received audio information using a natural language processing device and evaluates the degree of danger of the spoken content,

[0983] An analytical means for analyzing emotional states from video information using facial recognition calculations,

[0984] A generation means that determines the degree of risk based on harassment indicators and generates feedback information,

[0985] A notification means for notifying the user of the generated feedback information,

[0986] A storage means for saving analysis information and history,

[0987] A training method for retraining a machine learning model based on stored information,

[0988] A notification system using visual devices to detect signs of harassment in real time and provide feedback to staff,

[0989] A system that includes this.

[0990] (Claim 2)

[0991] The system according to claim 1, wherein the transmitted audio and video information is encrypted before transmission.

[0992] (Claim 3)

[0993] The system according to claim 1, wherein the accuracy of the analyzed harassment risk is improved by comparing it with past information. [Explanation of Symbols]

[0994] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. Acquisition means for acquiring acoustic and visual information, A transmission means for transmitting acquired acoustic and visual information via a communication protocol, An evaluation method that analyzes received acoustic information using a natural language processing engine and evaluates the danger of the content of the speech, An analytical means for analyzing emotional states from visual information using an facial expression recognition algorithm, A generation means that determines the degree of risk based on harassment indicators and generates feedback information, A notification means for notifying the user of the generated feedback information, A storage means for storing analysis information and historical information, A training method for retraining a machine learning model based on stored information, A means having the function of monitoring conversations within the home, detecting specific phrases to support smooth communication, and providing feedback, A system that includes this.

2. The system according to claim 1, wherein the transmitted acoustic information and visual information are encrypted before transmission.

3. The system according to claim 1, wherein the accuracy of the analyzed harassment risk is improved by comparing it with past information.

Images