system

Abstract

We provide the system. [Solution] A means for preprocessing data acquired from a data acquisition device, A means for generating a machine learning model that performs training using preprocessed data, A means for evaluating the generated machine learning model and retraining the model based on the evaluation results, A method for analyzing data in real time using a retrained model, A means of notifying the user of the analysis results, A system that includes this.

Description

Technical Field

[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 performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a conventional data analysis system, it has been difficult to efficiently perform consistent processing from data collection to analysis and result notification. Also, realizing real-time anomaly detection while maintaining the accuracy of the learning model has been a technical problem.

Means for Solving the Problems

[0005] This invention solves the above problems by providing a system that preprocesses data acquired from a data collection device, generates a machine learning model, and evaluates and retrains the model. Thereby, real-time data analysis is performed and the analysis results are notified to the user, realizing an improvement in the accuracy of anomaly detection.

[0006] A "data acquisition device" is a device consisting of hardware and software for acquiring data, and has the function of collecting data via various sensors and networks.

[0007] "Preprocessing" refers to the process of transforming collected data into an analyzable format, and includes operations such as noise reduction, standardization, and normalization.

[0008] A "machine learning model" is a collection of algorithms that learn from collected data and perform predictions and classifications, and includes neural networks and decision trees.

[0009] "Evaluation" is the process of measuring the performance of a machine learning model and confirming its accuracy and reliability, using metrics such as accuracy, recall, and precision.

[0010] "Retraining" refers to the process of retraining an existing machine learning model using additional data in order to improve or maintain its performance.

[0011] "Real-time analysis" is a process that performs analysis immediately upon data collection, and promptly extracts and provides the necessary information.

[0012] "User notifications" refer to the process of communicating analysis results and important information from the system to the user, and are provided in the form of alerts and reports. [Brief explanation of the drawing]

[0013] [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] It 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] It is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] It 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] It is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] It 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] It shows an emotion map to which a plurality of emotions are mapped. [Figure 10] It shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Embodiments for Carrying Out the Invention

[0014] 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.

[0015] First, the language used in the following description will be explained.

[0016] In the following embodiments, the labeled processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple 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.

[0017] In the following embodiments, the labeled RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0018] In the following embodiments, the labeled storage 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.

[0019] In the following embodiments, the labeled communication I / F (Interface) is an interface that includes a communication processor and an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark), and the like.

[0020] 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."

[0021] [First Embodiment]

[0022] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0023] 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.

[0024] 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).

[0025] 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.

[0026] 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.

[0027] 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.

[0028] 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.

[0029] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0030] 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.

[0031] 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.

[0032] 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.

[0033] 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".

[0034] To implement this invention, it is necessary to construct a system that integrates the processes of data collection, analysis, and result notification. In the embodiments of this invention, a server plays a central role, starting with collecting various types of data from data collection devices.

[0035] First, the server acquires data in real time from network cameras and sensors. This includes periodic polling and event-driven data collection methods. The collected data is stored in high-speed storage and then preprocessed, such as denoising and format conversion.

[0036] Next, the server performs analysis using machine learning models based on the pre-processed data. This process utilizes models specifically designed for tasks such as anomaly detection and classification. The accuracy and validity of the analysis results are verified by an evaluation function. While existing pre-trained models are used here, retraining is performed if the model's performance degrades due to the addition of new data. The latest dataset is used for retraining, enabling more accurate analysis.

[0037] Furthermore, the server processes the important information obtained from the analysis in real time and sends the results to the terminal. The terminal then presents this information to the user via a user interface. The user can use this interface to view detailed analysis results and take action as needed.

[0038] As a concrete example, let's consider an implementation for security purposes. The server collects video data, and when it detects abnormal behavior, it issues an alarm to the terminal. This alarm allows the user to respond in real time, enabling swift action such as dispatching security guards to the scene.

[0039] Another application example is a patient monitoring system in a medical facility. The server detects abnormalities from the patient's video feed and notifies medical staff via a terminal, facilitating a rapid medical response.

[0040] Thus, the present invention provides specific embodiments that enable high-precision real-time anomaly detection and immediate notification to users, thereby improving convenience and efficiency in various business areas.

[0041] The following describes the processing flow.

[0042] Step 1:

[0043] The server collects data from various sensors and network cameras. At this stage, real-time data acquisition is performed using streaming data polling and event listening. The collected data is temporarily stored in a database or cloud storage.

[0044] Step 2:

[0045] The server performs preprocessing on the collected data. For video data, it applies a noise reduction filter and standardizes the resolution. For text data, it performs tokenization and stop word removal, converting it into a data format suitable for analysis. This preprocessing improves the accuracy and efficiency of anomaly detection.

[0046] Step 3:

[0047] The server runs a machine learning model using pre-processed data. The model aims to detect abnormal behavior and specific events, employing a neural network-based approach. The model extracts features from the data and generates results based on pre-defined criteria.

[0048] Step 4:

[0049] The server evaluates the output of the machine learning model. Here, the accuracy of the results is verified using metrics such as precision, recall, and accuracy. If the model's performance falls below a certain threshold, a retraining process is triggered, and the model is retrained using additional datasets to update its parameters.

[0050] Step 5:

[0051] The server sends analysis results to the terminal. This includes providing information to the user through a user interface, in the form of real-time alerts and periodic reports.

[0052] Step 6:

[0053] The terminal displays alerts and reports received from the server on the user interface. Users can then take appropriate action immediately based on this information. For example, a user who receives an anomaly detection notification can dispatch response staff or provide feedback to the system.

[0054] (Example 1)

[0055] 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."

[0056] There is a need to process large amounts of data in real time while performing highly accurate anomaly detection. However, conventional technologies suffer from problems with real-time processing and accuracy due to the time required for data preprocessing and analysis. Therefore, a system is needed that can quickly and accurately detect anomalies and notify users.

[0057] 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.

[0058] In this invention, the server includes means for preprocessing information acquired from an information acquisition device, means for performing format conversion while removing noise, and means for notifying a human via a terminal of the analysis results. This enables rapid and highly accurate analysis of information and real-time notification of anomaly detection results.

[0059] An "information acquisition device" is a device, such as a network camera or sensor, that collects data in real time from the environment and circumstances.

[0060] "Preprocessing" is the process of preparing data for analysis by removing noise from acquired information and converting its format.

[0061] A "machine learning model" is a model that uses algorithms to learn from data and perform anomaly detection and classification.

[0062] "Noise reduction" is the process of removing unwanted signals and information in order to improve the accuracy of data analysis.

[0063] "Format conversion" is the process of changing the format of data in order to make it easier to analyze.

[0064] "Analysis results" refer to the results of analyzing data obtained by a machine learning model, and include anomaly detection and data classification.

[0065] A "terminal" is a device that displays analysis results notified from a server and provides an interface that allows humans to verify the information.

[0066] To implement this invention, a system integrating a data acquisition device, a server, a terminal, and a user interface is required. The server acquires data in real time from information acquisition devices such as network cameras and sensors. In this case, a polling method, which collects information at regular intervals, or an event-driven method can be used for data acquisition.

[0067] Next, the server saves the acquired data to high-speed storage. Since this saved data may contain noise, it undergoes noise reduction processing and further preprocessing to convert the data into the format required for analysis. Gaussian filters and other methods are used for noise reduction.

[0068] Subsequently, the server performs analysis using a machine learning model based on the pre-processed data. This model is specialized for anomaly detection and classification tasks and uses pre-trained data. The accuracy and validity of the analysis results are verified using an evaluation function. Based on the evaluation of the analysis results, the model is retrained as needed to improve accuracy by incorporating new data.

[0069] The information obtained through the analysis is transmitted to the terminal in real time by the server. The terminal then presents the analysis results to the user via a user interface. The user can use this interface to review the displayed information and take quick action as needed.

[0070] A concrete example is a security system. A server collects video data from local network cameras and, upon detecting abnormal activity, issues an alarm via a terminal. This allows users to view the footage in real time and quickly dispatch security personnel. This entire process aims to enable the system to perform rapid and highly accurate anomaly detection, prompting users to take immediate action.

[0071] An example of a prompt would be, "Design a system that uses a machine learning model to detect abnormal behavior for security purposes and notifies the user in real time."

[0072] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0073] Step 1:

[0074] The server acquires data in real time from information acquisition devices. The input is raw data from network cameras and sensors. This data is collected at regular intervals or based on event triggers. The output is raw data stored in high-speed storage. In this process, the API of each device is called to acquire data and perform initial data storage.

[0075] Step 2:

[0076] The server preprocesses the stored data. Specifically, it performs denoising and format conversion. The input is the stored raw data. A Gaussian filter is applied for denoising, and the data is converted into a format suitable for analysis. The output is the clean data after preprocessing.

[0077] Step 3:

[0078] The server runs a machine learning model using pre-processed data. The input is clean, pre-processed data. The data is analyzed by the trained model to perform anomaly detection and classification. The output is analytical information obtained as a result of whether or not anomalies are present and the classification results. The model is pre-trained on a sufficient dataset and is ready for immediate application.

[0079] Step 4:

[0080] The server evaluates the obtained analysis results. Indicators such as accuracy and F-score are used for evaluation to confirm performance. The input is the analysis results. If the evaluation reveals a decrease in model accuracy, the retraining process is initiated. The output is the evaluation score and a judgment on whether retraining is necessary.

[0081] Step 5:

[0082] The server sends the evaluated analysis results to the terminal. The input is the scrutinized analysis results. When sent to the terminal, the results are converted into a display format for the user interface. The output is the information presented to the user on the terminal.

[0083] Step 6:

[0084] The user checks the results on the terminal and takes action as needed. Alarms and detailed information are displayed on the terminal. User input (e.g., instructions to dispatch security personnel) is accepted and the corresponding action is taken. Output is the execution of user actions and the system's reaction.

[0085] (Application Example 1)

[0086] 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."

[0087] In modern society, with the increasing need for crime prevention and surveillance, there is a demand for effective systems that can quickly and accurately detect anomalies and notify users of that information in real time. However, existing technologies struggle to improve the accuracy of anomaly detection and achieve immediate notification. Furthermore, they lack the support functions necessary for users to respond quickly based on detected anomalies.

[0088] 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.

[0089] In this invention, the server includes means for preprocessing data acquired from a data collection device, means for generating a machine learning model that performs learning using the preprocessed data, and means for evaluating the generated machine learning model and retraining the model based on the evaluation results. This enables improved accuracy of real-time anomaly detection and rapid user notification.

[0090] A "data acquisition device" is a device, such as a sensor or network camera, that is installed to collect various numerical data and images.

[0091] "Preprocessing" refers to applying processes such as noise reduction and format conversion to the acquired data to prepare it for analysis.

[0092] A "machine learning model" is a mathematical algorithm that generates patterns from collected data to perform predictions and classifications.

[0093] "Retraining" is the process of retraining an existing machine learning model using newly acquired data to improve its performance.

[0094] "Real-time analysis" is a processing method that performs analysis immediately at the moment data is acquired, and obtains results instantly.

[0095] A "user interface" is a system that visually displays analysis results and provides a screen for users to take corresponding actions.

[0096] "Data communication technology" refers to network technology used to transmit analyzed information to devices in remote locations.

[0097] "Event information" refers to information about anomalies or specific conditions detected through analysis, and is subject to notification and recording.

[0098] In order to implement this invention, it is necessary to construct a system that combines a server, a data acquisition device, and a user terminal.

[0099] The server is built on an AWS® EC2 instance and receives data (video and sensor data) acquired from sensors and network cameras in real time via the network. The received data is streamed smoothly using Apache® Kafka. Pandas is used for data preprocessing, including noise reduction and format conversion. The data, prepared in this preprocessing stage, is then input into a machine learning model built using TENSORFLOW®. This model is designed to perform anomaly detection and normal behavior classification at high speed.

[0100] The server uses AWS SNS to send notifications to user terminals based on event information generated through analysis. These notifications are crucial for users to recognize anomalies in real time and take immediate action. Detailed analysis results are displayed on the user terminal, making it easier for users to identify abnormal conditions and take immediate action as needed.

[0101] As a concrete example, if applied to an office's nighttime surveillance system, when abnormal activity is detected, the user can receive a push notification on their smartphone stating, "Suspicious activity detected in the office. Please check the video." Tapping the notification allows them to view details of the abnormal activity and video footage before and after the incident.

[0102] Example of a prompt:

[0103] "Detect any abnormal behavior from the office's nighttime surveillance footage. If any activity is detected, determine whether it is related to normal work procedures."

[0104] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0105] Step 1:

[0106] The server receives data in real time from data collection devices such as sensors and network cameras. This data, which includes video and sensor data, is managed on an AWS EC2 instance. The input is raw data acquired from the data collection devices, and the output is unformatted, unprocessed data.

[0107] Step 2:

[0108] The server streams the received data using Apache Kafka and preprocesses it using Pandas. The data format is standardized and noise is removed to prepare it for analysis. At this stage, the input is the streamed raw data, and the output is preprocessed data with a consistent format.

[0109] Step 3:

[0110] The server inputs pre-processed data into a machine learning model using TensorFlow and performs anomaly detection. The generative AI model uses newly designed prompt statements to determine whether or not an anomaly is present. The input is pre-processed data, and the output is the analysis result. Here, it determines whether or not an anomaly was detected.

[0111] Step 4:

[0112] Based on the analysis results, the server uses AWS SNS to notify the user's device of event information. This is to immediately alert the user if an anomaly is detected. The input is the result of the anomaly detection, and the output is a push notification.

[0113] Step 5:

[0114] The terminal displays received notifications on the user interface, allowing the user to view detailed analysis results. This enables the user to identify abnormal activity and take immediate action. The input is notifications from the server, and the output is the information displayed on the user interface.

[0115] 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.

[0116] To implement this invention, it is necessary to combine a data analysis system with an emotion engine to recognize the user's emotions from the collected data and optimize the presentation of the analysis results. In the embodiments of this invention, a server plays a major role in integrating the processes of data collection, preprocessing, analysis, and emotion recognition.

[0117] First, the server collects data on user movements and facial expressions from network cameras and sensors. This data is stored in a database in real time and preprocessed, including noise reduction and standardization. The preprocessed data is then sent to a machine learning model, which includes an emotion recognition algorithm.

[0118] The server's machine learning model is designed to analyze data and identify specific behavioral patterns and user emotional states with high accuracy. The emotion engine identifies emotions hidden in actions and facial expressions and classifies them using emotion categories such as positive, negative, and neutral. The model's output undergoes an evaluation process to confirm its accuracy and reliability.

[0119] Based on the evaluation results, the server retrains the model. During the real-time analysis phase, the retrained model is used to generate analysis results. Furthermore, the emotion engine incorporates the user's emotional tendencies to optimize how the analysis results are presented. For example, if the user is feeling stressed, the analysis results are displayed in a simplified form, and actionable recommendations are provided.

[0120] The results are immediately sent to the device, allowing for user feedback. Users can review the analysis results and sentiment data through the user interface and take action based on them. The system is continuously improved based on the feedback.

[0121] As a concrete example, consider a customer service scenario. A server analyzes video data from customer interactions and uses an emotion engine to recognize the customer's emotions. As a result, the user (operator) can understand the customer's emotional state and select an appropriate response. In the entertainment field, content recommendations are customized according to the user's emotions.

[0122] This allows the system to provide insights tailored to the user's needs and emotional state, resulting in a more personalized experience.

[0123] The following describes the processing flow.

[0124] Step 1:

[0125] The server collects user movement, facial expressions, and voice data in real time from network cameras and sensors. This requires a system that monitors input from each sensor and saves it to a database at regular intervals.

[0126] Step 2:

[0127] The server preprocesses the collected raw data. For video data, a face detection algorithm is used to extract face regions, and a noise reduction filter is applied. Audio data undergoes noise filtering and spectral transformation, and is converted into a format suitable for analysis.

[0128] Step 3:

[0129] The server inputs pre-processed data into the emotion engine. The emotion engine uses a machine learning model to recognize emotions from the user's facial expressions and changes in voice, and classifies them into emotion categories such as positive, negative, and neutral.

[0130] Step 4:

[0131] The server personalizes the analysis results based on the user's emotions obtained by the emotion engine. At this stage, it determines how to present the data according to the user's emotional state, and adaptive processing is performed, such as omitting details or providing the data in a different format as needed.

[0132] Step 5:

[0133] The terminal displays analysis results and sentiment data from the server on the user interface. Based on this, the user can understand the situation and select appropriate actions. For example, if negative emotions are detected, the user can request a more detailed analysis or ask for assistance.

[0134] Step 6:

[0135] Users input feedback into their devices based on the information provided. This feedback is sent to the server and used to continuously improve the system. The server uses the collected feedback data to fine-tune and retrain the emotion engine, further improving the accuracy of the analysis.

[0136] (Example 2)

[0137] 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 as the "terminal".

[0138] In today's information-saturated society, there is a need to accurately analyze the information and emotional state of individual users and provide appropriate feedback. Existing systems struggle to respond quickly to changes in emotions, limiting their ability to optimize the user experience. To solve this problem, technology is needed that accurately recognizes the user's emotional state and provides analysis results in real time.

[0139] 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.

[0140] In this invention, the server includes means for preprocessing information collected from data acquisition devices, means for generating an artificial intelligence model that learns using the preprocessed information, and means for evaluating the generated artificial intelligence model and retraining the model based on the evaluation results. This makes it possible to analyze the user's emotional state and provide optimal feedback in real time in response to its changes.

[0141] A "data acquisition device" is a device that collects information related to the user's actions and facial expressions.

[0142] "Preprocessing" refers to processes such as noise reduction and data standardization that prepare collected information into a format suitable for analysis.

[0143] An "artificial intelligence model" is a program that uses a learning algorithm to analyze data and is used to identify a user's emotional state.

[0144] "Evaluation" is the process of confirming the analytical accuracy and reliability of the generated artificial intelligence model.

[0145] "Retraining" refers to the re-execution of training to improve the accuracy of an artificial intelligence model based on evaluation results.

[0146] "Real-time analysis" is an analysis method that processes acquired information immediately and provides results to the user.

[0147] "Emotional state" refers to a category of emotions identified from a user's actions and facial expressions, and indicates the user's psychological state.

[0148] "User terminal means" refers to devices and interfaces for providing users with analysis results and emotional data.

[0149] This invention provides a system that analyzes a user's movements and facial expressions and recognizes their emotional state in real time.

[0150] First, the server collects information related to the user's actions and facial expressions in real time through data acquisition devices. Network cameras and sensors are used as these data acquisition devices. This information is stored in a database. The stored information is then preprocessed, such as noise reduction and standardization, to prepare it for analysis.

[0151] The server then supplies the pre-processed information to the artificial intelligence model. This AI model uses a generative AI model and includes an emotion recognition algorithm. The model analyzes the user's information and identifies categories of emotions such as positive, negative, and neutral. The accuracy and reliability of these analysis results are verified through an evaluation process. Based on the evaluation results, the model is retrained as needed. Retraining improves the model's analytical capabilities and yields more accurate results.

[0152] The analysis results are sent to the device and provided to the user via a user interface. Through this interface, the user can review the analyzed information and sentiment data and decide on further actions. The interface presents the analysis results in an easily understandable format and recommends actionable steps as needed.

[0153] For example, in a customer service scenario, the server can analyze video data from customer interactions and use an emotion engine to recognize the customer's emotions. As a result, the operator can understand the customer's emotions and choose an appropriate response. In the entertainment field, it's also possible to recommend content based on the user's emotions.

[0154] Finally, an example of a prompt using a generative AI model is, "Analyze the customer's emotional state from their facial expression data and suggest a response for the operator." In this way, the system can provide a personalized experience that responds to the user's emotional state.

[0155] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0156] Step 1:

[0157] The server collects data on user movements and facial expressions from network cameras and sensors. Inputs include real-time image and sensor data. This data is first stored in a database. Specific operations include receiving the camera's video stream and temporarily saving it. The output of this step is raw data ready for preprocessing.

[0158] Step 2:

[0159] The server performs preprocessing on the collected data. The input is the raw data obtained in Step 1. Data quality is improved through processes such as noise reduction and data standardization. Specifically, unnecessary noise is filtered from the image data, and the data is converted into a format suitable for analysis. The output of this step is the preprocessed data.

[0160] Step 3:

[0161] The server inputs the preprocessed data into the machine learning model. This input is the preprocessed data output in step 2. The model uses a generative AI model to analyze the data. Specifically, it extracts features from the data and identifies the user's emotional state. The output of this step is the emotional information as a result of the analysis.

[0162] Step 4:

[0163] The server evaluates the analysis results and retrains the model as needed. The input is the analysis results obtained in step 3, and the server evaluates the model's accuracy based on these results. Specific actions include error calculation based on evaluation criteria and updating the dataset for retraining. The output of this step is the optimized model and its evaluation results.

[0164] Step 5:

[0165] The server sends the analysis results to the terminal and notifies the user. The input is the analysis and evaluation results obtained in steps 3 and 4. Specifically, the process includes converting the analysis results into a display format via the user interface and sending it to the terminal. The output of this step is the analysis result display received by the user.

[0166] Step 6:

[0167] The user reviews the analysis results and sentiment information via their device. The input is the analysis results sent in step 5. Based on this information, the user decides on an action and provides feedback to the system as needed. Specific actions include interpreting the results on the screen and considering countermeasures. The output of this step is the user's feedback and actions.

[0168] (Application Example 2)

[0169] 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".

[0170] Providing appropriate responses and suggestions that address residents' emotional states within the home environment is believed to contribute to reducing stress and improving their quality of life. However, conventional home systems have struggled to accurately recognize emotional states in real time and optimize interactions based on that information. This has resulted in problems in responding appropriately to residents' needs.

[0171] 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.

[0172] In this invention, the server includes means for preprocessing user behavior data and voice data acquired from a data collection device; means for generating a machine learning model that learns using the preprocessed data; means for analyzing the data in real time using the retrained model to identify the user's emotional state; and means for determining a response corresponding to the identified emotional state and notifying the user. This makes it possible to provide appropriate responses in real time that are in line with the user's emotions in a home environment.

[0173] A "data acquisition device" is a device consisting of hardware and software for acquiring user behavior data and voice data.

[0174] "Preprocessing" refers to the process of applying noise reduction, standardization, and other treatments to collected data to prepare it for analysis.

[0175] A "machine learning model" refers to an algorithm that uses pre-processed data to learn and recognize or predict specific patterns.

[0176] "Retraining" is the process of performing additional training to improve an existing machine learning model based on evaluation results.

[0177] "Real-time analysis" refers to the immediate processing of collected data to instantly identify the user's emotional state and other relevant information.

[0178] "Emotional state" refers to the psychological or emotional condition identified from the user's actions and tone of voice.

[0179] "Notification" refers to the act of communicating analysis results and response content to the user.

[0180] The system designed to realize this application is designed to understand the user's emotional state in real time within the home and provide appropriate responses and suggestions based on that understanding.

[0181] The server first acquires user behavior and voice data via a data collection device. This data is sent to the server in real time and undergoes preprocessing such as noise reduction and standardization. The preprocessed data is then input into a generative AI model for emotion recognition using machine learning frameworks such as TensorFlow or PyTorch. The model identifies the user's emotional state from the data and classifies it into categories such as positive, negative, and neutral.

[0182] The device makes appropriate suggestions to the user based on emotional state information received from the server. For example, if the device identifies that the user is stressed, it can respond by suggesting that the user play relaxing music.

[0183] Users view the analysis results through a user interface on their device and provide feedback on the suggestions offered. This feedback is sent to the server and used as part of the model's retraining process to improve the overall system accuracy.

[0184] For example, if a user comes home tired after a long day at work, the system can analyze the user's tone of voice and facial expression and suggest, "Shall I play some music to help you relax?"

[0185] An example of a prompt message is: "Design a robot application that offers stress-reducing suggestions within the home based on the user's emotions. For example, it should include a feature that suggests relaxation music when the user is feeling stressed."

[0186] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0187] Step 1:

[0188] The server acquires user motion data and voice data in real time from data acquisition devices. Since the data may be raw and contain noise, preprocessing is performed immediately after data acquisition. The server receives data streams from cameras and microphones as input and generates standardized, pre-processed data as output, with noise removed.

[0189] Step 2:

[0190] The server inputs preprocessed data into a generative AI model for emotion recognition. Here, TensorFlow or PyTorch is used to identify specific emotion patterns from the data. Using preprocessed data as input, the output is a label indicating the user's emotional state (positive, negative, neutral, etc.). Algorithms such as multilayer neural networks are used in this process.

[0191] Step 3:

[0192] The server determines what suggestions or responses to offer the user based on the identified emotional state. The input is the recognized emotional label, and the output is the corresponding response or suggestion. At this stage, for example, if the user is feeling stressed, an action such as suggesting playing relaxation music might be selected.

[0193] Step 4:

[0194] The terminal notifies the user of the suggestions received from the server. A user interface is utilized here, conveying the suggestions or responses to the user visually or audibly. The input is the suggestions from the server, and the output is the user receiving visualized or audible information.

[0195] Step 5:

[0196] Users provide feedback on the suggestions offered through the terminal's user interface. The input is user feedback information, and the output is data sent to the server. This feedback is used in subsequent relearning processes.

[0197] Step 6:

[0198] The server retrains the generative AI model based on accumulated feedback data to improve the system's accuracy. The input is feedback data and past model states, and the output is the updated generative AI model. Retraining is performed periodically to improve the model's performance.

[0199] 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.

[0200] 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.

[0201] 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.

[0202] [Second Embodiment]

[0203] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0204] 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.

[0205] 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).

[0206] 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.

[0207] 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.

[0208] 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).

[0209] 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.

[0210] 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.

[0211] 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.

[0212] 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.

[0213] 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.

[0214] 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".

[0215] To implement this invention, it is necessary to construct a system that integrates the processes of data collection, analysis, and result notification. In the embodiments of this invention, a server plays a central role, starting with collecting various types of data from data collection devices.

[0216] First, the server acquires data in real time from network cameras and sensors. This includes periodic polling and event-driven data collection methods. The collected data is stored in high-speed storage and then preprocessed, such as denoising and format conversion.

[0217] Next, the server performs analysis using machine learning models based on the pre-processed data. This process utilizes models specifically designed for tasks such as anomaly detection and classification. The accuracy and validity of the analysis results are verified by an evaluation function. While existing pre-trained models are used here, retraining is performed if the model's performance degrades due to the addition of new data. The latest dataset is used for retraining, enabling more accurate analysis.

[0218] Furthermore, the server processes the important information obtained from the analysis in real time and sends the results to the terminal. The terminal then presents this information to the user via a user interface. The user can use this interface to view detailed analysis results and take action as needed.

[0219] As a concrete example, let's consider an implementation for security purposes. The server collects video data, and when it detects abnormal behavior, it issues an alarm to the terminal. This alarm allows the user to respond in real time, enabling swift action such as dispatching security guards to the scene.

[0220] Another application example is a patient monitoring system in a medical facility. The server detects abnormalities from the patient's video feed and notifies medical staff via a terminal, facilitating a rapid medical response.

[0221] Thus, the present invention provides specific embodiments that enable high-precision real-time anomaly detection and immediate notification to users, thereby improving convenience and efficiency in various business areas.

[0222] The following describes the processing flow.

[0223] Step 1:

[0224] The server collects data from various sensors and network cameras. At this stage, real-time data acquisition is performed using streaming data polling and event listening. The collected data is temporarily stored in a database or cloud storage.

[0225] Step 2:

[0226] The server performs preprocessing on the collected data. For video data, it applies a noise reduction filter and standardizes the resolution. For text data, it performs tokenization and stop word removal, converting it into a data format suitable for analysis. This preprocessing improves the accuracy and efficiency of anomaly detection.

[0227] Step 3:

[0228] The server runs a machine learning model using pre-processed data. The model aims to detect abnormal behavior and specific events, employing a neural network-based approach. The model extracts features from the data and generates results based on pre-defined criteria.

[0229] Step 4:

[0230] The server evaluates the output of the machine learning model. Here, the accuracy of the results is verified using metrics such as precision, recall, and accuracy. If the model's performance falls below a certain threshold, a retraining process is triggered, and the model is retrained using additional datasets to update its parameters.

[0231] Step 5:

[0232] The server sends analysis results to the terminal. This includes providing information to the user through a user interface, in the form of real-time alerts and periodic reports.

[0233] Step 6:

[0234] The terminal displays alerts and reports received from the server on the user interface. Users can then take appropriate action immediately based on this information. For example, a user who receives an anomaly detection notification can dispatch response staff or provide feedback to the system.

[0235] (Example 1)

[0236] 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."

[0237] There is a need to process large amounts of data in real time while performing highly accurate anomaly detection. However, conventional technologies suffer from problems with real-time processing and accuracy due to the time required for data preprocessing and analysis. Therefore, a system is needed that can quickly and accurately detect anomalies and notify users.

[0238] 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.

[0239] In this invention, the server includes means for preprocessing information acquired from an information acquisition device, means for performing format conversion while removing noise, and means for notifying a human via a terminal of the analysis results. This enables rapid and highly accurate analysis of information and real-time notification of anomaly detection results.

[0240] An "information acquisition device" is a device, such as a network camera or sensor, that collects data in real time from the environment and circumstances.

[0241] "Preprocessing" is the process of preparing data for analysis by removing noise from acquired information and converting its format.

[0242] A "machine learning model" is a model that uses algorithms to learn from data and perform anomaly detection and classification.

[0243] "Noise reduction" is the process of removing unwanted signals and information in order to improve the accuracy of data analysis.

[0244] "Format conversion" is the process of changing the format of data in order to make it easier to analyze.

[0245] "Analysis results" refer to the results of analyzing data obtained by a machine learning model, and include anomaly detection and data classification.

[0246] A "terminal" is a device that displays analysis results notified from a server and provides an interface that allows humans to verify the information.

[0247] To implement this invention, a system integrating a data acquisition device, a server, a terminal, and a user interface is required. The server acquires data in real time from information acquisition devices such as network cameras and sensors. In this case, a polling method, which collects information at regular intervals, or an event-driven method can be used for data acquisition.

[0248] Next, the server saves the acquired data to high-speed storage. Since this saved data may contain noise, it undergoes noise reduction processing and further preprocessing to convert the data into the format required for analysis. Gaussian filters and other methods are used for noise reduction.

[0249] Subsequently, the server performs analysis using a machine learning model based on the pre-processed data. This model is specialized for anomaly detection and classification tasks and uses pre-trained data. The accuracy and validity of the analysis results are verified using an evaluation function. Based on the evaluation of the analysis results, the model is retrained as needed to improve accuracy by incorporating new data.

[0250] The information obtained through the analysis is transmitted to the terminal in real time by the server. The terminal then presents the analysis results to the user via a user interface. The user can use this interface to review the displayed information and take quick action as needed.

[0251] A concrete example is a security system. A server collects video data from local network cameras and, upon detecting abnormal activity, issues an alarm via a terminal. This allows users to view the footage in real time and quickly dispatch security personnel. This entire process aims to enable the system to perform rapid and highly accurate anomaly detection, prompting users to take immediate action.

[0252] An example of a prompt would be, "Design a system that uses a machine learning model to detect abnormal behavior for security purposes and notifies the user in real time."

[0253] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0254] Step 1:

[0255] The server acquires data in real time from information acquisition devices. The input is raw data from network cameras and sensors. This data is collected at regular intervals or based on event triggers. The output is raw data stored in high-speed storage. In this process, the API of each device is called to acquire data and perform initial data storage.

[0256] Step 2:

[0257] The server preprocesses the stored data. Specifically, it performs denoising and format conversion. The input is the stored raw data. A Gaussian filter is applied for denoising, and the data is converted into a format suitable for analysis. The output is the clean data after preprocessing.

[0258] Step 3:

[0259] The server runs a machine learning model using pre-processed data. The input is clean, pre-processed data. The data is analyzed by the trained model to perform anomaly detection and classification. The output is analytical information obtained as a result of whether or not anomalies are present and the classification results. The model is pre-trained on a sufficient dataset and is ready for immediate application.

[0260] Step 4:

[0261] The server evaluates the obtained analysis results. Indicators such as accuracy and F-score are used for evaluation to confirm performance. The input is the analysis results. If the evaluation reveals a decrease in model accuracy, the retraining process is initiated. The output is the evaluation score and a judgment on whether retraining is necessary.

[0262] Step 5:

[0263] The server sends the evaluated analysis results to the terminal. The input is the scrutinized analysis results. When sent to the terminal, the results are converted into a display format for the user interface. The output is the information presented to the user on the terminal.

[0264] Step 6:

[0265] The user checks the results on the terminal and takes action as needed. Alarms and detailed information are displayed on the terminal. User input (e.g., instructions to dispatch security personnel) is accepted and the corresponding action is taken. Output is the execution of user actions and the system's reaction.

[0266] (Application Example 1)

[0267] 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."

[0268] In modern society, with the increasing need for crime prevention and surveillance, there is a demand for effective systems that can quickly and accurately detect anomalies and notify users of that information in real time. However, existing technologies struggle to improve the accuracy of anomaly detection and achieve immediate notification. Furthermore, they lack the support functions necessary for users to respond quickly based on detected anomalies.

[0269] 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.

[0270] In this invention, the server includes means for preprocessing data acquired from a data collection device, means for generating a machine learning model that performs learning using the preprocessed data, and means for evaluating the generated machine learning model and retraining the model based on the evaluation results. This enables improved accuracy of real-time anomaly detection and rapid user notification.

[0271] A "data acquisition device" is a device, such as a sensor or network camera, that is installed to collect various numerical data and images.

[0272] "Preprocessing" refers to applying processes such as noise reduction and format conversion to the acquired data to prepare it for analysis.

[0273] A "machine learning model" is a mathematical algorithm that generates patterns from collected data to perform predictions and classifications.

[0274] "Retraining" is the process of retraining an existing machine learning model using newly acquired data to improve its performance.

[0275] "Real-time analysis" is a processing method that performs analysis immediately at the moment data is acquired, and obtains results instantly.

[0276] A "user interface" is a system that visually displays analysis results and provides a screen for users to take corresponding actions.

[0277] "Data communication technology" refers to network technology used to transmit analyzed information to devices in remote locations.

[0278] "Event information" refers to information about anomalies or specific conditions detected through analysis, and is subject to notification and recording.

[0279] In order to implement this invention, it is necessary to construct a system that combines a server, a data acquisition device, and a user terminal.

[0280] The server is built on an AWS EC2 instance and receives data (video and sensor data) acquired from sensors and network cameras in real time via the network. The received data is streamed smoothly using Apache Kafka. Pandas is used for data preprocessing, such as noise reduction and format conversion. The data, prepared in this preprocessing stage, is then input into a machine learning model built using TensorFlow. This model is designed to perform anomaly detection and normal behavior classification at high speed.

[0281] The server uses AWS SNS to send notifications to user terminals based on event information generated through analysis. These notifications are crucial for users to recognize anomalies in real time and take immediate action. Detailed analysis results are displayed on the user terminal, making it easier for users to identify abnormal conditions and take immediate action as needed.

[0282] As a specific example, when applied to an office night surveillance system, if abnormal activities are detected, the user can receive a push notification on their smartphone saying, "Suspicious activities have been detected in the office. Please check the video." Tapping on the notification allows the user to view details of the abnormal activities and the video before and after it.

[0283] Example of prompt text:

[0284] "Please detect abnormal behavior from the office night surveillance video. If there is activity, please determine whether the behavior is due to normal business operations."

[0285] The flow of the specific process in Application Example 1 will be described using FIG. 12.

[0286] Step 1:

[0287] The server receives data in real time from data collection devices such as sensors and network cameras. This data is video data and sensor data and is managed on an AWS EC2 instance. The input is raw data obtained from the data collection devices, and the output is the unformatted data as it is.

[0288] Step 2:

[0289] The server streams the received data using Apache Kafka and performs preprocessing using Pandas. By unifying the data format and removing noise, it is arranged in a form suitable for analysis. The input at this stage is the streamed raw data, and the output is the preprocessed data with a formatted appearance.

[0290] Step 3:

[0291] The server inputs pre-processed data into a machine learning model using TensorFlow and performs anomaly detection. The generative AI model uses newly designed prompt statements to determine whether or not an anomaly is present. The input is pre-processed data, and the output is the analysis result. Here, it determines whether or not an anomaly was detected.

[0292] Step 4:

[0293] Based on the analysis results, the server uses AWS SNS to notify the user's device of event information. This is to immediately alert the user if an anomaly is detected. The input is the result of the anomaly detection, and the output is a push notification.

[0294] Step 5:

[0295] The terminal displays received notifications on the user interface, allowing the user to view detailed analysis results. This enables the user to identify abnormal activity and take immediate action. The input is notifications from the server, and the output is the information displayed on the user interface.

[0296] 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.

[0297] To implement this invention, it is necessary to combine a data analysis system with an emotion engine to recognize the user's emotions from the collected data and optimize the presentation of the analysis results. In the embodiments of this invention, a server plays a major role in integrating the processes of data collection, preprocessing, analysis, and emotion recognition.

[0298] First, the server collects data on user movements and facial expressions from network cameras and sensors. This data is stored in a database in real time and preprocessed, including noise reduction and standardization. The preprocessed data is then sent to a machine learning model, which includes an emotion recognition algorithm.

[0299] The server's machine learning model is designed to analyze data and identify specific behavioral patterns and user emotional states with high accuracy. The emotion engine identifies emotions hidden in actions and facial expressions and classifies them using emotion categories such as positive, negative, and neutral. The model's output undergoes an evaluation process to confirm its accuracy and reliability.

[0300] Based on the evaluation results, the server retrains the model. During the real-time analysis phase, the retrained model is used to generate analysis results. Furthermore, the emotion engine incorporates the user's emotional tendencies to optimize how the analysis results are presented. For example, if the user is feeling stressed, the analysis results are displayed in a simplified form, and actionable recommendations are provided.

[0301] The results are immediately sent to the device, allowing for user feedback. Users can review the analysis results and sentiment data through the user interface and take action based on them. The system is continuously improved based on the feedback.

[0302] As a concrete example, consider a customer service scenario. A server analyzes video data from customer interactions and uses an emotion engine to recognize the customer's emotions. As a result, the user (operator) can understand the customer's emotional state and select an appropriate response. In the entertainment field, content recommendations are customized according to the user's emotions.

[0303] This allows the system to provide insights tailored to the user's needs and emotional state, resulting in a more personalized experience.

[0304] The processing flow will be described below.

[0305] Step 1:

[0306] The server collects the user's motion, expression, and voice data in real time from network cameras and sensors. For this, a mechanism for monitoring the input from each sensor and storing it in a database at regular intervals is required.

[0307] Step 2:

[0308] The server performs preprocessing on the collected raw data. For video data, a face detection algorithm is used to extract the face region, and a noise removal filter is applied. The voice data is subjected to noise filtering and spectral conversion and is converted into a format suitable for analysis.

[0309] Step 3:

[0310] The server inputs the preprocessed data into the emotion engine. The emotion engine uses a machine learning model to recognize emotions from the changes in the user's facial expressions and voice and classifies them into emotion categories such as positive, negative, and neutral.

[0311] Step 4:

[0312] The server personalizes the analysis results based on the user's emotions obtained by the emotion engine. At this stage, an adaptation process is performed to determine a data presentation method according to the user's emotional state and to omit details or provide them in a different format as necessary.

[0313] Step 5:

[0314] The terminal displays analysis results and sentiment data from the server on the user interface. Based on this, the user can understand the situation and select appropriate actions. For example, if negative emotions are detected, the user can request a more detailed analysis or ask for assistance.

[0315] Step 6:

[0316] Users input feedback into their devices based on the information provided. This feedback is sent to the server and used to continuously improve the system. The server uses the collected feedback data to fine-tune and retrain the emotion engine, further improving the accuracy of the analysis.

[0317] (Example 2)

[0318] 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".

[0319] In today's information-saturated society, there is a need to accurately analyze the information and emotional state of individual users and provide appropriate feedback. Existing systems struggle to respond quickly to changes in emotions, limiting their ability to optimize the user experience. To solve this problem, technology is needed that accurately recognizes the user's emotional state and provides analysis results in real time.

[0320] 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.

[0321] In this invention, the server includes means for preprocessing information collected from data acquisition devices, means for generating an artificial intelligence model that learns using the preprocessed information, and means for evaluating the generated artificial intelligence model and retraining the model based on the evaluation results. This makes it possible to analyze the user's emotional state and provide optimal feedback in real time in response to its changes.

[0322] A "data acquisition device" is a device that collects information related to the user's actions and facial expressions.

[0323] "Preprocessing" refers to processes such as noise reduction and data standardization that prepare collected information into a format suitable for analysis.

[0324] An "artificial intelligence model" is a program that uses a learning algorithm to analyze data and is used to identify a user's emotional state.

[0325] "Evaluation" is the process of confirming the analytical accuracy and reliability of the generated artificial intelligence model.

[0326] "Retraining" refers to the re-execution of training to improve the accuracy of an artificial intelligence model based on evaluation results.

[0327] "Real-time analysis" is an analysis method that processes acquired information immediately and provides results to the user.

[0328] "Emotional state" refers to a category of emotions identified from a user's actions and facial expressions, and indicates the user's psychological state.

[0329] "User terminal means" refers to devices and interfaces for providing users with analysis results and emotional data.

[0330] This invention provides a system that analyzes a user's movements and facial expressions and recognizes their emotional state in real time.

[0331] First, the server collects information related to the user's actions and facial expressions in real time through data acquisition devices. Network cameras and sensors are used as these data acquisition devices. This information is stored in a database. The stored information is then preprocessed, such as noise reduction and standardization, to prepare it for analysis.

[0332] The server then supplies the pre-processed information to the artificial intelligence model. This AI model uses a generative AI model and includes an emotion recognition algorithm. The model analyzes the user's information and identifies categories of emotions such as positive, negative, and neutral. The accuracy and reliability of these analysis results are verified through an evaluation process. Based on the evaluation results, the model is retrained as needed. Retraining improves the model's analytical capabilities and yields more accurate results.

[0333] The analysis results are sent to the device and provided to the user via a user interface. Through this interface, the user can review the analyzed information and sentiment data and decide on further actions. The interface presents the analysis results in an easily understandable format and recommends actionable steps as needed.

[0334] For example, in a customer service scenario, the server can analyze video data from customer interactions and use an emotion engine to recognize the customer's emotions. As a result, the operator can understand the customer's emotions and choose an appropriate response. In the entertainment field, it's also possible to recommend content based on the user's emotions.

[0335] Finally, an example of a prompt using a generative AI model is, "Analyze the customer's emotional state from their facial expression data and suggest a response for the operator." In this way, the system can provide a personalized experience that responds to the user's emotional state.

[0336] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0337] Step 1:

[0338] The server collects data on user movements and facial expressions from network cameras and sensors. Inputs include real-time image and sensor data. This data is first stored in a database. Specific operations include receiving the camera's video stream and temporarily saving it. The output of this step is raw data ready for preprocessing.

[0339] Step 2:

[0340] The server performs preprocessing on the collected data. The input is the raw data obtained in Step 1. Data quality is improved through processes such as noise reduction and data standardization. Specifically, unnecessary noise is filtered from the image data, and the data is converted into a format suitable for analysis. The output of this step is the preprocessed data.

[0341] Step 3:

[0342] The server inputs the preprocessed data into the machine learning model. This input is the preprocessed data output in step 2. The model uses a generative AI model to analyze the data. Specifically, it extracts features from the data and identifies the user's emotional state. The output of this step is the emotional information as a result of the analysis.

[0343] Step 4:

[0344] The server evaluates the analysis results and retrains the model as needed. The input is the analysis results obtained in step 3, and the server evaluates the model's accuracy based on these results. Specific actions include error calculation based on evaluation criteria and updating the dataset for retraining. The output of this step is the optimized model and its evaluation results.

[0345] Step 5:

[0346] The server sends the analysis results to the terminal and notifies the user. The input is the analysis and evaluation results obtained in steps 3 and 4. Specifically, the process includes converting the analysis results into a display format via the user interface and sending it to the terminal. The output of this step is the analysis result display received by the user.

[0347] Step 6:

[0348] The user reviews the analysis results and sentiment information via their device. The input is the analysis results sent in step 5. Based on this information, the user decides on an action and provides feedback to the system as needed. Specific actions include interpreting the results on the screen and considering countermeasures. The output of this step is the user's feedback and actions.

[0349] (Application Example 2)

[0350] 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 will be referred to as the "terminal."

[0351] Providing appropriate responses and suggestions that address residents' emotional states within the home environment is believed to contribute to reducing stress and improving their quality of life. However, conventional home systems have struggled to accurately recognize emotional states in real time and optimize interactions based on that information. This has resulted in problems in responding appropriately to residents' needs.

[0352] 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.

[0353] In this invention, the server includes means for preprocessing user behavior data and voice data acquired from a data collection device; means for generating a machine learning model that learns using the preprocessed data; means for analyzing the data in real time using the retrained model to identify the user's emotional state; and means for determining a response corresponding to the identified emotional state and notifying the user. This makes it possible to provide appropriate responses in real time that are in line with the user's emotions in a home environment.

[0354] A "data acquisition device" is a device consisting of hardware and software for acquiring user behavior data and voice data.

[0355] "Preprocessing" refers to the process of applying noise reduction, standardization, and other treatments to collected data to prepare it for analysis.

[0356] A "machine learning model" refers to an algorithm that uses pre-processed data to learn and recognize or predict specific patterns.

[0357] "Retraining" is the process of performing additional training to improve an existing machine learning model based on evaluation results.

[0358] "Real-time analysis" refers to the immediate processing of collected data to instantly identify the user's emotional state and other relevant information.

[0359] "Emotional state" refers to the psychological or emotional condition identified from the user's actions and tone of voice.

[0360] "Notification" refers to the act of communicating analysis results and response content to the user.

[0361] The system designed to realize this application is designed to understand the user's emotional state in real time within the home and provide appropriate responses and suggestions based on that understanding.

[0362] The server first acquires user behavior and voice data via a data collection device. This data is sent to the server in real time and undergoes preprocessing such as noise reduction and standardization. The preprocessed data is then input into a generative AI model for emotion recognition using machine learning frameworks such as TensorFlow or PyTorch. The model identifies the user's emotional state from the data and classifies it into categories such as positive, negative, and neutral.

[0363] The device makes appropriate suggestions to the user based on emotional state information received from the server. For example, if the device identifies that the user is stressed, it can respond by suggesting that the user play relaxing music.

[0364] Users view the analysis results through a user interface on their device and provide feedback on the suggestions offered. This feedback is sent to the server and used as part of the model's retraining process to improve the overall system accuracy.

[0365] For example, if a user comes home tired after a long day at work, the system can analyze the user's tone of voice and facial expression and suggest, "Shall I play some music to help you relax?"

[0366] An example of a prompt message is: "Design a robot application that offers stress-reducing suggestions within the home based on the user's emotions. For example, it should include a feature that suggests relaxation music when the user is feeling stressed."

[0367] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0368] Step 1:

[0369] The server acquires user motion data and voice data in real time from data acquisition devices. Since the data may be raw and contain noise, preprocessing is performed immediately after data acquisition. The server receives data streams from cameras and microphones as input and generates standardized, pre-processed data as output, with noise removed.

[0370] Step 2:

[0371] The server inputs preprocessed data into a generative AI model for emotion recognition. Here, TensorFlow or PyTorch is used to identify specific emotion patterns from the data. Using preprocessed data as input, the output is a label indicating the user's emotional state (positive, negative, neutral, etc.). Algorithms such as multilayer neural networks are used in this process.

[0372] Step 3:

[0373] The server determines what suggestions or responses to offer the user based on the identified emotional state. The input is the recognized emotional label, and the output is the corresponding response or suggestion. At this stage, for example, if the user is feeling stressed, an action such as suggesting playing relaxation music might be selected.

[0374] Step 4:

[0375] The terminal notifies the user of the suggestions received from the server. A user interface is utilized here, conveying the suggestions or responses to the user visually or audibly. The input is the suggestions from the server, and the output is the user receiving visualized or audible information.

[0376] Step 5:

[0377] Users provide feedback on the suggestions offered through the terminal's user interface. The input is user feedback information, and the output is data sent to the server. This feedback is used in subsequent relearning processes.

[0378] Step 6:

[0379] The server retrains the generative AI model based on accumulated feedback data to improve the system's accuracy. The input is feedback data and past model states, and the output is the updated generative AI model. Retraining is performed periodically to improve the model's performance.

[0380] 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.

[0381] 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.

[0382] 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.

[0383] [Third Embodiment]

[0384] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0385] 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.

[0386] 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).

[0387] 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.

[0388] 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.

[0389] 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).

[0390] 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.

[0391] 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.

[0392] 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.

[0393] 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.

[0394] 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.

[0395] 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".

[0396] To implement this invention, it is necessary to construct a system that integrates the processes of data collection, analysis, and result notification. In the embodiments of this invention, a server plays a central role, starting with collecting various types of data from data collection devices.

[0397] First, the server acquires data in real time from network cameras and sensors. This includes periodic polling and event-driven data collection methods. The collected data is stored in high-speed storage and then preprocessed, such as denoising and format conversion.

[0398] Next, the server performs analysis using machine learning models based on the pre-processed data. This process utilizes models specifically designed for tasks such as anomaly detection and classification. The accuracy and validity of the analysis results are verified by an evaluation function. While existing pre-trained models are used here, retraining is performed if the model's performance degrades due to the addition of new data. The latest dataset is used for retraining, enabling more accurate analysis.

[0399] Furthermore, the server processes the important information obtained from the analysis in real time and sends the results to the terminal. The terminal then presents this information to the user via a user interface. The user can use this interface to view detailed analysis results and take action as needed.

[0400] As a concrete example, let's consider an implementation for security purposes. The server collects video data, and when it detects abnormal behavior, it issues an alarm to the terminal. This alarm allows the user to respond in real time, enabling swift action such as dispatching security guards to the scene.

[0401] Another application example is a patient monitoring system in a medical facility. The server detects abnormalities from the patient's video feed and notifies medical staff via a terminal, facilitating a rapid medical response.

[0402] Thus, the present invention provides specific embodiments that enable high-precision real-time anomaly detection and immediate notification to users, thereby improving convenience and efficiency in various business areas.

[0403] The following describes the processing flow.

[0404] Step 1:

[0405] The server collects data from various sensors and network cameras. At this stage, real-time data acquisition is performed using streaming data polling and event listening. The collected data is temporarily stored in a database or cloud storage.

[0406] Step 2:

[0407] The server performs preprocessing on the collected data. For video data, it applies a noise reduction filter and standardizes the resolution. For text data, it performs tokenization and stop word removal, converting it into a data format suitable for analysis. This preprocessing improves the accuracy and efficiency of anomaly detection.

[0408] Step 3:

[0409] The server runs a machine learning model using pre-processed data. The model aims to detect abnormal behavior and specific events, employing a neural network-based approach. The model extracts features from the data and generates results based on pre-defined criteria.

[0410] Step 4:

[0411] The server evaluates the output of the machine learning model. Here, the accuracy of the results is verified using metrics such as precision, recall, and accuracy. If the model's performance falls below a certain threshold, a retraining process is triggered, and the model is retrained using additional datasets to update its parameters.

[0412] Step 5:

[0413] The server sends analysis results to the terminal. This includes providing information to the user through a user interface, in the form of real-time alerts and periodic reports.

[0414] Step 6:

[0415] The terminal displays alerts and reports received from the server on the user interface. Users can then take appropriate action immediately based on this information. For example, a user who receives an anomaly detection notification can dispatch response staff or provide feedback to the system.

[0416] (Example 1)

[0417] 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."

[0418] There is a need to process large amounts of data in real time while performing highly accurate anomaly detection. However, conventional technologies suffer from problems with real-time processing and accuracy due to the time required for data preprocessing and analysis. Therefore, a system is needed that can quickly and accurately detect anomalies and notify users.

[0419] 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.

[0420] In this invention, the server includes means for preprocessing information acquired from an information acquisition device, means for performing format conversion while removing noise, and means for notifying a human via a terminal of the analysis results. This enables rapid and highly accurate analysis of information and real-time notification of anomaly detection results.

[0421] An "information acquisition device" is a device, such as a network camera or sensor, that collects data in real time from the environment and circumstances.

[0422] "Preprocessing" is the process of preparing data for analysis by removing noise from acquired information and converting its format.

[0423] A "machine learning model" is a model that uses algorithms to learn from data and perform anomaly detection and classification.

[0424] "Noise reduction" is the process of removing unwanted signals and information in order to improve the accuracy of data analysis.

[0425] "Format conversion" is the process of changing the format of data in order to make it easier to analyze.

[0426] "Analysis results" refer to the results of analyzing data obtained by a machine learning model, and include anomaly detection and data classification.

[0427] A "terminal" is a device that displays analysis results notified from a server and provides an interface that allows humans to verify the information.

[0428] To implement this invention, a system integrating a data acquisition device, a server, a terminal, and a user interface is required. The server acquires data in real time from information acquisition devices such as network cameras and sensors. In this case, a polling method, which collects information at regular intervals, or an event-driven method can be used for data acquisition.

[0429] Next, the server saves the acquired data to high-speed storage. Since this saved data may contain noise, it undergoes noise reduction processing and further preprocessing to convert the data into the format required for analysis. Gaussian filters and other methods are used for noise reduction.

[0430] Subsequently, the server performs analysis using a machine learning model based on the pre-processed data. This model is specialized for anomaly detection and classification tasks and uses pre-trained data. The accuracy and validity of the analysis results are verified using an evaluation function. Based on the evaluation of the analysis results, the model is retrained as needed to improve accuracy by incorporating new data.

[0431] The information obtained through the analysis is transmitted to the terminal in real time by the server. The terminal then presents the analysis results to the user via a user interface. The user can use this interface to review the displayed information and take quick action as needed.

[0432] A concrete example is a security system. A server collects video data from local network cameras and, upon detecting abnormal activity, issues an alarm via a terminal. This allows users to view the footage in real time and quickly dispatch security personnel. This entire process aims to enable the system to perform rapid and highly accurate anomaly detection, prompting users to take immediate action.

[0433] An example of a prompt would be, "Design a system that uses a machine learning model to detect abnormal behavior for security purposes and notifies the user in real time."

[0434] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0435] Step 1:

[0436] The server acquires data in real time from information acquisition devices. The input is raw data from network cameras and sensors. This data is collected at regular intervals or based on event triggers. The output is raw data stored in high-speed storage. In this process, the API of each device is called to acquire data and perform initial data storage.

[0437] Step 2:

[0438] The server preprocesses the stored data. Specifically, it performs denoising and format conversion. The input is the stored raw data. A Gaussian filter is applied for denoising, and the data is converted into a format suitable for analysis. The output is the clean data after preprocessing.

[0439] Step 3:

[0440] The server runs a machine learning model using pre-processed data. The input is clean, pre-processed data. The data is analyzed by the trained model to perform anomaly detection and classification. The output is analytical information obtained as a result of whether or not anomalies are present and the classification results. The model is pre-trained on a sufficient dataset and is ready for immediate application.

[0441] Step 4:

[0442] The server evaluates the obtained analysis results. Indicators such as accuracy and F-score are used for evaluation to confirm performance. The input is the analysis results. If the evaluation reveals a decrease in model accuracy, the retraining process is initiated. The output is the evaluation score and a judgment on whether retraining is necessary.

[0443] Step 5:

[0444] The server sends the evaluated analysis results to the terminal. The input is the scrutinized analysis results. When sent to the terminal, the results are converted into a display format for the user interface. The output is the information presented to the user on the terminal.

[0445] Step 6:

[0446] The user checks the results on the terminal and takes action as needed. Alarms and detailed information are displayed on the terminal. User input (e.g., instructions to dispatch security personnel) is accepted and the corresponding action is taken. Output is the execution of user actions and the system's reaction.

[0447] (Application Example 1)

[0448] 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."

[0449] In modern society, with the increasing need for crime prevention and surveillance, there is a demand for effective systems that can quickly and accurately detect anomalies and notify users of that information in real time. However, existing technologies struggle to improve the accuracy of anomaly detection and achieve immediate notification. Furthermore, they lack the support functions necessary for users to respond quickly based on detected anomalies.

[0450] 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.

[0451] In this invention, the server includes means for preprocessing data acquired from a data collection device, means for generating a machine learning model that performs learning using the preprocessed data, and means for evaluating the generated machine learning model and retraining the model based on the evaluation results. This enables improved accuracy of real-time anomaly detection and rapid user notification.

[0452] A "data acquisition device" is a device, such as a sensor or network camera, that is installed to collect various numerical data and images.

[0453] "Preprocessing" refers to applying processes such as noise reduction and format conversion to the acquired data to prepare it for analysis.

[0454] A "machine learning model" is a mathematical algorithm that generates patterns from collected data to perform predictions and classifications.

[0455] "Retraining" is the process of retraining an existing machine learning model using newly acquired data to improve its performance.

[0456] "Real-time analysis" is a processing method that performs analysis immediately at the moment data is acquired, and obtains results instantly.

[0457] A "user interface" is a system that visually displays analysis results and provides a screen for users to take corresponding actions.

[0458] "Data communication technology" refers to network technology used to transmit analyzed information to devices in remote locations.

[0459] "Event information" refers to information about anomalies or specific conditions detected through analysis, and is subject to notification and recording.

[0460] In order to implement this invention, it is necessary to construct a system that combines a server, a data acquisition device, and a user terminal.

[0461] The server is built on an AWS EC2 instance and receives data (video and sensor data) acquired from sensors and network cameras in real time via the network. The received data is streamed smoothly using Apache Kafka. Pandas is used for data preprocessing, such as noise reduction and format conversion. The data, prepared in this preprocessing stage, is then input into a machine learning model built using TensorFlow. This model is designed to perform anomaly detection and normal behavior classification at high speed.

[0462] The server uses AWS SNS to send notifications to user terminals based on event information generated through analysis. These notifications are crucial for users to recognize anomalies in real time and take immediate action. Detailed analysis results are displayed on the user terminal, making it easier for users to identify abnormal conditions and take immediate action as needed.

[0463] As a concrete example, if applied to an office's nighttime surveillance system, when abnormal activity is detected, the user can receive a push notification on their smartphone stating, "Suspicious activity detected in the office. Please check the video." Tapping the notification allows them to view details of the abnormal activity and video footage before and after the incident.

[0464] Example of a prompt:

[0465] "Detect any abnormal behavior from the office's nighttime surveillance footage. If any activity is detected, determine whether it is related to normal work procedures."

[0466] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0467] Step 1:

[0468] The server receives data in real time from data collection devices such as sensors and network cameras. This data, which includes video and sensor data, is managed on an AWS EC2 instance. The input is raw data acquired from the data collection devices, and the output is unformatted, unprocessed data.

[0469] Step 2:

[0470] The server streams the received data using Apache Kafka and preprocesses it using Pandas. The data format is standardized and noise is removed to prepare it for analysis. At this stage, the input is the streamed raw data, and the output is preprocessed data with a consistent format.

[0471] Step 3:

[0472] The server inputs pre-processed data into a machine learning model using TensorFlow and performs anomaly detection. The generative AI model uses newly designed prompt statements to determine whether or not an anomaly is present. The input is pre-processed data, and the output is the analysis result. Here, it determines whether or not an anomaly was detected.

[0473] Step 4:

[0474] Based on the analysis results, the server uses AWS SNS to notify the user's device of event information. This is to immediately alert the user if an anomaly is detected. The input is the result of the anomaly detection, and the output is a push notification.

[0475] Step 5:

[0476] The terminal displays received notifications on the user interface, allowing the user to view detailed analysis results. This enables the user to identify abnormal activity and take immediate action. The input is notifications from the server, and the output is the information displayed on the user interface.

[0477] 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.

[0478] To implement this invention, it is necessary to combine a data analysis system with an emotion engine to recognize the user's emotions from the collected data and optimize the presentation of the analysis results. In the embodiments of this invention, a server plays a major role in integrating the processes of data collection, preprocessing, analysis, and emotion recognition.

[0479] First, the server collects data on user movements and facial expressions from network cameras and sensors. This data is stored in a database in real time and preprocessed, including noise reduction and standardization. The preprocessed data is then sent to a machine learning model, which includes an emotion recognition algorithm.

[0480] The server's machine learning model is designed to analyze data and identify specific behavioral patterns and user emotional states with high accuracy. The emotion engine identifies emotions hidden in actions and facial expressions and classifies them using emotion categories such as positive, negative, and neutral. The model's output undergoes an evaluation process to confirm its accuracy and reliability.

[0481] Based on the evaluation results, the server retrains the model. During the real-time analysis phase, the retrained model is used to generate analysis results. Furthermore, the emotion engine incorporates the user's emotional tendencies to optimize how the analysis results are presented. For example, if the user is feeling stressed, the analysis results are displayed in a simplified form, and actionable recommendations are provided.

[0482] The results are immediately sent to the device, allowing for user feedback. Users can review the analysis results and sentiment data through the user interface and take action based on them. The system is continuously improved based on the feedback.

[0483] As a concrete example, consider a customer service scenario. A server analyzes video data from customer interactions and uses an emotion engine to recognize the customer's emotions. As a result, the user (operator) can understand the customer's emotional state and select an appropriate response. In the entertainment field, content recommendations are customized according to the user's emotions.

[0484] This allows the system to provide insights tailored to the user's needs and emotional state, resulting in a more personalized experience.

[0485] The following describes the processing flow.

[0486] Step 1:

[0487] The server collects user movement, facial expressions, and voice data in real time from network cameras and sensors. This requires a system that monitors input from each sensor and saves it to a database at regular intervals.

[0488] Step 2:

[0489] The server preprocesses the collected raw data. For video data, a face detection algorithm is used to extract face regions, and a noise reduction filter is applied. Audio data undergoes noise filtering and spectral transformation, and is converted into a format suitable for analysis.

[0490] Step 3:

[0491] The server inputs pre-processed data into the emotion engine. The emotion engine uses a machine learning model to recognize emotions from the user's facial expressions and changes in voice, and classifies them into emotion categories such as positive, negative, and neutral.

[0492] Step 4:

[0493] The server personalizes the analysis results based on the user's emotions obtained by the emotion engine. At this stage, it determines how to present the data according to the user's emotional state, and adaptive processing is performed, such as omitting details or providing the data in a different format as needed.

[0494] Step 5:

[0495] The terminal displays analysis results and sentiment data from the server on the user interface. Based on this, the user can understand the situation and select appropriate actions. For example, if negative emotions are detected, the user can request a more detailed analysis or ask for assistance.

[0496] Step 6:

[0497] Users input feedback into their devices based on the information provided. This feedback is sent to the server and used to continuously improve the system. The server uses the collected feedback data to fine-tune and retrain the emotion engine, further improving the accuracy of the analysis.

[0498] (Example 2)

[0499] 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."

[0500] In today's information-saturated society, there is a need to accurately analyze the information and emotional state of individual users and provide appropriate feedback. Existing systems struggle to respond quickly to changes in emotions, limiting their ability to optimize the user experience. To solve this problem, technology is needed that accurately recognizes the user's emotional state and provides analysis results in real time.

[0501] 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.

[0502] In this invention, the server includes means for preprocessing information collected from data acquisition devices, means for generating an artificial intelligence model that learns using the preprocessed information, and means for evaluating the generated artificial intelligence model and retraining the model based on the evaluation results. This makes it possible to analyze the user's emotional state and provide optimal feedback in real time in response to its changes.

[0503] A "data acquisition device" is a device that collects information related to the user's actions and facial expressions.

[0504] "Preprocessing" refers to processes such as noise reduction and data standardization that prepare collected information into a format suitable for analysis.

[0505] An "artificial intelligence model" is a program that uses a learning algorithm to analyze data and is used to identify a user's emotional state.

[0506] "Evaluation" is the process of confirming the analytical accuracy and reliability of the generated artificial intelligence model.

[0507] "Retraining" refers to the re-execution of training to improve the accuracy of an artificial intelligence model based on evaluation results.

[0508] "Real-time analysis" is an analysis method that processes acquired information immediately and provides results to the user.

[0509] "Emotional state" refers to a category of emotions identified from a user's actions and facial expressions, and indicates the user's psychological state.

[0510] "User terminal means" refers to devices and interfaces for providing users with analysis results and emotional data.

[0511] This invention provides a system that analyzes a user's movements and facial expressions and recognizes their emotional state in real time.

[0512] First, the server collects information related to the user's actions and facial expressions in real time through data acquisition devices. Network cameras and sensors are used as these data acquisition devices. This information is stored in a database. The stored information is then preprocessed, such as noise reduction and standardization, to prepare it for analysis.

[0513] The server then supplies the pre-processed information to the artificial intelligence model. This AI model uses a generative AI model and includes an emotion recognition algorithm. The model analyzes the user's information and identifies categories of emotions such as positive, negative, and neutral. The accuracy and reliability of these analysis results are verified through an evaluation process. Based on the evaluation results, the model is retrained as needed. Retraining improves the model's analytical capabilities and yields more accurate results.

[0514] The analysis results are sent to the device and provided to the user via a user interface. Through this interface, the user can review the analyzed information and sentiment data and decide on further actions. The interface presents the analysis results in an easily understandable format and recommends actionable steps as needed.

[0515] For example, in a customer service scenario, the server can analyze video data from customer interactions and use an emotion engine to recognize the customer's emotions. As a result, the operator can understand the customer's emotions and choose an appropriate response. In the entertainment field, it's also possible to recommend content based on the user's emotions.

[0516] Finally, an example of a prompt using a generative AI model is, "Analyze the customer's emotional state from their facial expression data and suggest a response for the operator." In this way, the system can provide a personalized experience that responds to the user's emotional state.

[0517] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0518] Step 1:

[0519] The server collects data on user movements and facial expressions from network cameras and sensors. Inputs include real-time image and sensor data. This data is first stored in a database. Specific operations include receiving the camera's video stream and temporarily saving it. The output of this step is raw data ready for preprocessing.

[0520] Step 2:

[0521] The server performs preprocessing on the collected data. The input is the raw data obtained in Step 1. Data quality is improved through processes such as noise reduction and data standardization. Specifically, unnecessary noise is filtered from the image data, and the data is converted into a format suitable for analysis. The output of this step is the preprocessed data.

[0522] Step 3:

[0523] The server inputs the preprocessed data into the machine learning model. This input is the preprocessed data output in step 2. The model uses a generative AI model to analyze the data. Specifically, it extracts features from the data and identifies the user's emotional state. The output of this step is the emotional information as a result of the analysis.

[0524] Step 4:

[0525] The server evaluates the analysis results and retrains the model as needed. The input is the analysis results obtained in step 3, and the server evaluates the model's accuracy based on these results. Specific actions include error calculation based on evaluation criteria and updating the dataset for retraining. The output of this step is the optimized model and its evaluation results.

[0526] Step 5:

[0527] The server sends the analysis results to the terminal and notifies the user. The input is the analysis and evaluation results obtained in steps 3 and 4. Specifically, the process includes converting the analysis results into a display format via the user interface and sending it to the terminal. The output of this step is the analysis result display received by the user.

[0528] Step 6:

[0529] The user reviews the analysis results and sentiment information via their device. The input is the analysis results sent in step 5. Based on this information, the user decides on an action and provides feedback to the system as needed. Specific actions include interpreting the results on the screen and considering countermeasures. The output of this step is the user's feedback and actions.

[0530] (Application Example 2)

[0531] 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."

[0532] Providing appropriate responses and suggestions that address residents' emotional states within the home environment is believed to contribute to reducing stress and improving their quality of life. However, conventional home systems have struggled to accurately recognize emotional states in real time and optimize interactions based on that information. This has resulted in problems in responding appropriately to residents' needs.

[0533] 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.

[0534] In this invention, the server includes means for preprocessing user behavior data and voice data acquired from a data collection device; means for generating a machine learning model that learns using the preprocessed data; means for analyzing the data in real time using the retrained model to identify the user's emotional state; and means for determining a response corresponding to the identified emotional state and notifying the user. This makes it possible to provide appropriate responses in real time that are in line with the user's emotions in a home environment.

[0535] A "data acquisition device" is a device consisting of hardware and software for acquiring user behavior data and voice data.

[0536] "Preprocessing" refers to the process of applying noise reduction, standardization, and other treatments to collected data to prepare it for analysis.

[0537] A "machine learning model" refers to an algorithm that uses pre-processed data to learn and recognize or predict specific patterns.

[0538] "Retraining" is the process of performing additional training to improve an existing machine learning model based on evaluation results.

[0539] "Real-time analysis" refers to the immediate processing of collected data to instantly identify the user's emotional state and other relevant information.

[0540] "Emotional state" refers to the psychological or emotional condition identified from the user's actions and tone of voice.

[0541] "Notification" refers to the act of communicating analysis results and response content to the user.

[0542] The system designed to realize this application is designed to understand the user's emotional state in real time within the home and provide appropriate responses and suggestions based on that understanding.

[0543] The server first acquires user behavior and voice data via a data collection device. This data is sent to the server in real time and undergoes preprocessing such as noise reduction and standardization. The preprocessed data is then input into a generative AI model for emotion recognition using machine learning frameworks such as TensorFlow or PyTorch. The model identifies the user's emotional state from the data and classifies it into categories such as positive, negative, and neutral.

[0544] The device makes appropriate suggestions to the user based on emotional state information received from the server. For example, if the device identifies that the user is stressed, it can respond by suggesting that the user play relaxing music.

[0545] Users view the analysis results through a user interface on their device and provide feedback on the suggestions offered. This feedback is sent to the server and used as part of the model's retraining process to improve the overall system accuracy.

[0546] For example, if a user comes home tired after a long day at work, the system can analyze the user's tone of voice and facial expression and suggest, "Shall I play some music to help you relax?"

[0547] An example of a prompt message is: "Design a robot application that offers stress-reducing suggestions within the home based on the user's emotions. For example, it should include a feature that suggests relaxation music when the user is feeling stressed."

[0548] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0549] Step 1:

[0550] The server acquires user motion data and voice data in real time from data acquisition devices. Since the data may be raw and contain noise, preprocessing is performed immediately after data acquisition. The server receives data streams from cameras and microphones as input and generates standardized, pre-processed data as output, with noise removed.

[0551] Step 2:

[0552] The server inputs preprocessed data into a generative AI model for emotion recognition. Here, TensorFlow or PyTorch is used to identify specific emotion patterns from the data. Using preprocessed data as input, the output is a label indicating the user's emotional state (positive, negative, neutral, etc.). Algorithms such as multilayer neural networks are used in this process.

[0553] Step 3:

[0554] The server determines what suggestions or responses to offer the user based on the identified emotional state. The input is the recognized emotional label, and the output is the corresponding response or suggestion. At this stage, for example, if the user is feeling stressed, an action such as suggesting playing relaxation music might be selected.

[0555] Step 4:

[0556] The terminal notifies the user of the suggestions received from the server. A user interface is utilized here, conveying the suggestions or responses to the user visually or audibly. The input is the suggestions from the server, and the output is the user receiving visualized or audible information.

[0557] Step 5:

[0558] Users provide feedback on the suggestions offered through the terminal's user interface. The input is user feedback information, and the output is data sent to the server. This feedback is used in subsequent relearning processes.

[0559] Step 6:

[0560] The server retrains the generative AI model based on accumulated feedback data to improve the system's accuracy. The input is feedback data and past model states, and the output is the updated generative AI model. Retraining is performed periodically to improve the model's performance.

[0561] 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.

[0562] 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.

[0563] 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.

[0564] [Fourth Embodiment]

[0565] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0566] 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.

[0567] 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).

[0568] 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.

[0569] 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.

[0570] 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).

[0571] 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.

[0572] 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.

[0573] 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.

[0574] 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.

[0575] 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.

[0576] 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.

[0577] 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".

[0578] To implement this invention, it is necessary to construct a system that integrates the processes of data collection, analysis, and result notification. In the embodiments of this invention, a server plays a central role, starting with collecting various types of data from data collection devices.

[0579] First, the server acquires data in real time from network cameras and sensors. This includes periodic polling and event-driven data collection methods. The collected data is stored in high-speed storage and then preprocessed, such as denoising and format conversion.

[0580] Next, the server performs analysis using machine learning models based on the pre-processed data. This process utilizes models specifically designed for tasks such as anomaly detection and classification. The accuracy and validity of the analysis results are verified by an evaluation function. While existing pre-trained models are used here, retraining is performed if the model's performance degrades due to the addition of new data. The latest dataset is used for retraining, enabling more accurate analysis.

[0581] Furthermore, the server processes the important information obtained from the analysis in real time and sends the results to the terminal. The terminal then presents this information to the user via a user interface. The user can use this interface to view detailed analysis results and take action as needed.

[0582] As a concrete example, let's consider an implementation for security purposes. The server collects video data, and when it detects abnormal behavior, it issues an alarm to the terminal. This alarm allows the user to respond in real time, enabling swift action such as dispatching security guards to the scene.

[0583] Another application example is a patient monitoring system in a medical facility. The server detects abnormalities from the patient's video feed and notifies medical staff via a terminal, facilitating a rapid medical response.

[0584] Thus, the present invention provides specific embodiments that enable high-precision real-time anomaly detection and immediate notification to users, thereby improving convenience and efficiency in various business areas.

[0585] The following describes the processing flow.

[0586] Step 1:

[0587] The server collects data from various sensors and network cameras. At this stage, real-time data acquisition is performed using streaming data polling and event listening. The collected data is temporarily stored in a database or cloud storage.

[0588] Step 2:

[0589] The server performs preprocessing on the collected data. For video data, it applies a noise reduction filter and standardizes the resolution. For text data, it performs tokenization and stop word removal, converting it into a data format suitable for analysis. This preprocessing improves the accuracy and efficiency of anomaly detection.

[0590] Step 3:

[0591] The server runs a machine learning model using pre-processed data. The model aims to detect abnormal behavior and specific events, employing a neural network-based approach. The model extracts features from the data and generates results based on pre-defined criteria.

[0592] Step 4:

[0593] The server evaluates the output of the machine learning model. Here, the accuracy of the results is verified using metrics such as precision, recall, and accuracy. If the model's performance falls below a certain threshold, a retraining process is triggered, and the model is retrained using additional datasets to update its parameters.

[0594] Step 5:

[0595] The server sends analysis results to the terminal. This includes providing information to the user through a user interface, in the form of real-time alerts and periodic reports.

[0596] Step 6:

[0597] The terminal displays alerts and reports received from the server on the user interface. Users can then take appropriate action immediately based on this information. For example, a user who receives an anomaly detection notification can dispatch response staff or provide feedback to the system.

[0598] (Example 1)

[0599] 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".

[0600] There is a need to process large amounts of data in real time while performing highly accurate anomaly detection. However, conventional technologies suffer from problems with real-time processing and accuracy due to the time required for data preprocessing and analysis. Therefore, a system is needed that can quickly and accurately detect anomalies and notify users.

[0601] 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.

[0602] In this invention, the server includes means for preprocessing information acquired from an information acquisition device, means for performing format conversion while removing noise, and means for notifying a human via a terminal of the analysis results. This enables rapid and highly accurate analysis of information and real-time notification of anomaly detection results.

[0603] An "information acquisition device" is a device, such as a network camera or sensor, that collects data in real time from the environment and circumstances.

[0604] "Preprocessing" is the process of preparing data for analysis by removing noise from acquired information and converting its format.

[0605] A "machine learning model" is a model that uses algorithms to learn from data and perform anomaly detection and classification.

[0606] "Noise reduction" is the process of removing unwanted signals and information in order to improve the accuracy of data analysis.

[0607] "Format conversion" is the process of changing the format of data in order to make it easier to analyze.

[0608] "Analysis results" refer to the results of analyzing data obtained by a machine learning model, and include anomaly detection and data classification.

[0609] A "terminal" is a device that displays analysis results notified from a server and provides an interface that allows humans to verify the information.

[0610] To implement this invention, a system integrating a data acquisition device, a server, a terminal, and a user interface is required. The server acquires data in real time from information acquisition devices such as network cameras and sensors. In this case, a polling method, which collects information at regular intervals, or an event-driven method can be used for data acquisition.

[0611] Next, the server saves the acquired data to high-speed storage. Since this saved data may contain noise, it undergoes noise reduction processing and further preprocessing to convert the data into the format required for analysis. Gaussian filters and other methods are used for noise reduction.

[0612] Subsequently, the server performs analysis using a machine learning model based on the pre-processed data. This model is specialized for anomaly detection and classification tasks and uses pre-trained data. The accuracy and validity of the analysis results are verified using an evaluation function. Based on the evaluation of the analysis results, the model is retrained as needed to improve accuracy by incorporating new data.

[0613] The information obtained through the analysis is transmitted to the terminal in real time by the server. The terminal then presents the analysis results to the user via a user interface. The user can use this interface to review the displayed information and take quick action as needed.

[0614] A concrete example is a security system. A server collects video data from local network cameras and, upon detecting abnormal activity, issues an alarm via a terminal. This allows users to view the footage in real time and quickly dispatch security personnel. This entire process aims to enable the system to perform rapid and highly accurate anomaly detection, prompting users to take immediate action.

[0615] An example of a prompt would be, "Design a system that uses a machine learning model to detect abnormal behavior for security purposes and notifies the user in real time."

[0616] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0617] Step 1:

[0618] The server acquires data in real time from information acquisition devices. The input is raw data from network cameras and sensors. This data is collected at regular intervals or based on event triggers. The output is raw data stored in high-speed storage. In this process, the API of each device is called to acquire data and perform initial data storage.

[0619] Step 2:

[0620] The server preprocesses the stored data. Specifically, it performs denoising and format conversion. The input is the stored raw data. A Gaussian filter is applied for denoising, and the data is converted into a format suitable for analysis. The output is the clean data after preprocessing.

[0621] Step 3:

[0622] The server runs a machine learning model using pre-processed data. The input is clean, pre-processed data. The data is analyzed by the trained model to perform anomaly detection and classification. The output is analytical information obtained as a result of whether or not anomalies are present and the classification results. The model is pre-trained on a sufficient dataset and is ready for immediate application.

[0623] Step 4:

[0624] The server evaluates the obtained analysis results. Indicators such as accuracy and F-score are used for evaluation to confirm performance. The input is the analysis results. If the evaluation reveals a decrease in model accuracy, the retraining process is initiated. The output is the evaluation score and a judgment on whether retraining is necessary.

[0625] Step 5:

[0626] The server sends the evaluated analysis results to the terminal. The input is the scrutinized analysis results. When sent to the terminal, the results are converted into a display format for the user interface. The output is the information presented to the user on the terminal.

[0627] Step 6:

[0628] The user checks the results on the terminal and takes action as needed. Alarms and detailed information are displayed on the terminal. User input (e.g., instructions to dispatch security personnel) is accepted and the corresponding action is taken. Output is the execution of user actions and the system's reaction.

[0629] (Application Example 1)

[0630] 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".

[0631] In modern society, with the increasing need for crime prevention and surveillance, there is a demand for effective systems that can quickly and accurately detect anomalies and notify users of that information in real time. However, existing technologies struggle to improve the accuracy of anomaly detection and achieve immediate notification. Furthermore, they lack the support functions necessary for users to respond quickly based on detected anomalies.

[0632] 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.

[0633] In this invention, the server includes means for preprocessing data acquired from a data collection device, means for generating a machine learning model that performs learning using the preprocessed data, and means for evaluating the generated machine learning model and retraining the model based on the evaluation results. This enables improved accuracy of real-time anomaly detection and rapid user notification.

[0634] A "data acquisition device" is a device, such as a sensor or network camera, that is installed to collect various numerical data and images.

[0635] "Preprocessing" refers to applying processes such as noise reduction and format conversion to the acquired data to prepare it for analysis.

[0636] A "machine learning model" is a mathematical algorithm that generates patterns from collected data to perform predictions and classifications.

[0637] "Retraining" is the process of retraining an existing machine learning model using newly acquired data to improve its performance.

[0638] "Real-time analysis" is a processing method that performs analysis immediately at the moment data is acquired, and obtains results instantly.

[0639] A "user interface" is a system that visually displays analysis results and provides a screen for users to take corresponding actions.

[0640] "Data communication technology" refers to network technology used to transmit analyzed information to devices in remote locations.

[0641] "Event information" refers to information about anomalies or specific conditions detected through analysis, and is subject to notification and recording.

[0642] In order to implement this invention, it is necessary to construct a system that combines a server, a data acquisition device, and a user terminal.

[0643] The server is built on an AWS EC2 instance and receives data (video and sensor data) acquired from sensors and network cameras in real time via the network. The received data is streamed smoothly using Apache Kafka. Pandas is used for data preprocessing, such as noise reduction and format conversion. The data, prepared in this preprocessing stage, is then input into a machine learning model built using TensorFlow. This model is designed to perform anomaly detection and normal behavior classification at high speed.

[0644] The server uses AWS SNS to send notifications to user terminals based on event information generated through analysis. These notifications are crucial for users to recognize anomalies in real time and take immediate action. Detailed analysis results are displayed on the user terminal, making it easier for users to identify abnormal conditions and take immediate action as needed.

[0645] As a concrete example, if applied to an office's nighttime surveillance system, when abnormal activity is detected, the user can receive a push notification on their smartphone stating, "Suspicious activity detected in the office. Please check the video." Tapping the notification allows them to view details of the abnormal activity and video footage before and after the incident.

[0646] Example of a prompt:

[0647] "Detect any abnormal behavior from the office's nighttime surveillance footage. If any activity is detected, determine whether it is related to normal work procedures."

[0648] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0649] Step 1:

[0650] The server receives data in real time from data collection devices such as sensors and network cameras. This data, which includes video and sensor data, is managed on an AWS EC2 instance. The input is raw data acquired from the data collection devices, and the output is unformatted, unprocessed data.

[0651] Step 2:

[0652] The server streams the received data using Apache Kafka and preprocesses it using Pandas. The data format is standardized and noise is removed to prepare it for analysis. At this stage, the input is the streamed raw data, and the output is preprocessed data with a consistent format.

[0653] Step 3:

[0654] The server inputs pre-processed data into a machine learning model using TensorFlow and performs anomaly detection. The generative AI model uses newly designed prompt statements to determine whether or not an anomaly is present. The input is pre-processed data, and the output is the analysis result. Here, it determines whether or not an anomaly was detected.

[0655] Step 4:

[0656] Based on the analysis results, the server uses AWS SNS to notify the user's device of event information. This is to immediately alert the user if an anomaly is detected. The input is the result of the anomaly detection, and the output is a push notification.

[0657] Step 5:

[0658] The terminal displays received notifications on the user interface, allowing the user to view detailed analysis results. This enables the user to identify abnormal activity and take immediate action. The input is notifications from the server, and the output is the information displayed on the user interface.

[0659] 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.

[0660] To implement this invention, it is necessary to combine a data analysis system with an emotion engine to recognize the user's emotions from the collected data and optimize the presentation of the analysis results. In the embodiments of this invention, a server plays a major role in integrating the processes of data collection, preprocessing, analysis, and emotion recognition.

[0661] First, the server collects data on user movements and facial expressions from network cameras and sensors. This data is stored in a database in real time and preprocessed, including noise reduction and standardization. The preprocessed data is then sent to a machine learning model, which includes an emotion recognition algorithm.

[0662] The server's machine learning model is designed to analyze data and identify specific behavioral patterns and user emotional states with high accuracy. The emotion engine identifies emotions hidden in actions and facial expressions and classifies them using emotion categories such as positive, negative, and neutral. The model's output undergoes an evaluation process to confirm its accuracy and reliability.

[0663] Based on the evaluation results, the server retrains the model. During the real-time analysis phase, the retrained model is used to generate analysis results. Furthermore, the emotion engine incorporates the user's emotional tendencies to optimize how the analysis results are presented. For example, if the user is feeling stressed, the analysis results are displayed in a simplified form, and actionable recommendations are provided.

[0664] The results are immediately sent to the device, allowing for user feedback. Users can review the analysis results and sentiment data through the user interface and take action based on them. The system is continuously improved based on the feedback.

[0665] As a concrete example, consider a customer service scenario. A server analyzes video data from customer interactions and uses an emotion engine to recognize the customer's emotions. As a result, the user (operator) can understand the customer's emotional state and select an appropriate response. In the entertainment field, content recommendations are customized according to the user's emotions.

[0666] This allows the system to provide insights tailored to the user's needs and emotional state, resulting in a more personalized experience.

[0667] The following describes the processing flow.

[0668] Step 1:

[0669] The server collects user movement, facial expressions, and voice data in real time from network cameras and sensors. This requires a system that monitors input from each sensor and saves it to a database at regular intervals.

[0670] Step 2:

[0671] The server preprocesses the collected raw data. For video data, a face detection algorithm is used to extract face regions, and a noise reduction filter is applied. Audio data undergoes noise filtering and spectral transformation, and is converted into a format suitable for analysis.

[0672] Step 3:

[0673] The server inputs pre-processed data into the emotion engine. The emotion engine uses a machine learning model to recognize emotions from the user's facial expressions and changes in voice, and classifies them into emotion categories such as positive, negative, and neutral.

[0674] Step 4:

[0675] The server personalizes the analysis results based on the user's emotions obtained by the emotion engine. At this stage, it determines how to present the data according to the user's emotional state, and adaptive processing is performed, such as omitting details or providing the data in a different format as needed.

[0676] Step 5:

[0677] The terminal displays analysis results and sentiment data from the server on the user interface. Based on this, the user can understand the situation and select appropriate actions. For example, if negative emotions are detected, the user can request a more detailed analysis or ask for assistance.

[0678] Step 6:

[0679] Users input feedback into their devices based on the information provided. This feedback is sent to the server and used to continuously improve the system. The server uses the collected feedback data to fine-tune and retrain the emotion engine, further improving the accuracy of the analysis.

[0680] (Example 2)

[0681] 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".

[0682] In today's information-saturated society, there is a need to accurately analyze the information and emotional state of individual users and provide appropriate feedback. Existing systems struggle to respond quickly to changes in emotions, limiting their ability to optimize the user experience. To solve this problem, technology is needed that accurately recognizes the user's emotional state and provides analysis results in real time.

[0683] 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.

[0684] In this invention, the server includes means for preprocessing information collected from data acquisition devices, means for generating an artificial intelligence model that learns using the preprocessed information, and means for evaluating the generated artificial intelligence model and retraining the model based on the evaluation results. This makes it possible to analyze the user's emotional state and provide optimal feedback in real time in response to its changes.

[0685] A "data acquisition device" is a device that collects information related to the user's actions and facial expressions.

[0686] "Preprocessing" refers to processes such as noise reduction and data standardization that prepare collected information into a format suitable for analysis.

[0687] An "artificial intelligence model" is a program that uses a learning algorithm to analyze data and is used to identify a user's emotional state.

[0688] "Evaluation" is the process of confirming the analytical accuracy and reliability of the generated artificial intelligence model.

[0689] "Retraining" refers to the re-execution of training to improve the accuracy of an artificial intelligence model based on evaluation results.

[0690] "Real-time analysis" is an analysis method that processes acquired information immediately and provides results to the user.

[0691] "Emotional state" refers to a category of emotions identified from a user's actions and facial expressions, and indicates the user's psychological state.

[0692] "User terminal means" refers to devices and interfaces for providing users with analysis results and emotional data.

[0693] This invention provides a system that analyzes a user's movements and facial expressions and recognizes their emotional state in real time.

[0694] First, the server collects information related to the user's actions and facial expressions in real time through data acquisition devices. Network cameras and sensors are used as these data acquisition devices. This information is stored in a database. The stored information is then preprocessed, such as noise reduction and standardization, to prepare it for analysis.

[0695] The server then supplies the pre-processed information to the artificial intelligence model. This AI model uses a generative AI model and includes an emotion recognition algorithm. The model analyzes the user's information and identifies categories of emotions such as positive, negative, and neutral. The accuracy and reliability of these analysis results are verified through an evaluation process. Based on the evaluation results, the model is retrained as needed. Retraining improves the model's analytical capabilities and yields more accurate results.

[0696] The analysis results are sent to the device and provided to the user via a user interface. Through this interface, the user can review the analyzed information and sentiment data and decide on further actions. The interface presents the analysis results in an easily understandable format and recommends actionable steps as needed.

[0697] For example, in a customer service scenario, the server can analyze video data from customer interactions and use an emotion engine to recognize the customer's emotions. As a result, the operator can understand the customer's emotions and choose an appropriate response. In the entertainment field, it's also possible to recommend content based on the user's emotions.

[0698] Finally, an example of a prompt using a generative AI model is, "Analyze the customer's emotional state from their facial expression data and suggest a response for the operator." In this way, the system can provide a personalized experience that responds to the user's emotional state.

[0699] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0700] Step 1:

[0701] The server collects data on user movements and facial expressions from network cameras and sensors. Inputs include real-time image and sensor data. This data is first stored in a database. Specific operations include receiving the camera's video stream and temporarily saving it. The output of this step is raw data ready for preprocessing.

[0702] Step 2:

[0703] The server performs preprocessing on the collected data. The input is the raw data obtained in Step 1. Data quality is improved through processes such as noise reduction and data standardization. Specifically, unnecessary noise is filtered from the image data, and the data is converted into a format suitable for analysis. The output of this step is the preprocessed data.

[0704] Step 3:

[0705] The server inputs the preprocessed data into the machine learning model. This input is the preprocessed data output in step 2. The model uses a generative AI model to analyze the data. Specifically, it extracts features from the data and identifies the user's emotional state. The output of this step is the emotional information as a result of the analysis.

[0706] Step 4:

[0707] The server evaluates the analysis results and retrains the model as needed. The input is the analysis results obtained in step 3, and the server evaluates the model's accuracy based on these results. Specific actions include error calculation based on evaluation criteria and updating the dataset for retraining. The output of this step is the optimized model and its evaluation results.

[0708] Step 5:

[0709] The server sends the analysis results to the terminal and notifies the user. The input is the analysis and evaluation results obtained in steps 3 and 4. Specifically, the process includes converting the analysis results into a display format via the user interface and sending it to the terminal. The output of this step is the analysis result display received by the user.

[0710] Step 6:

[0711] The user reviews the analysis results and sentiment information via their device. The input is the analysis results sent in step 5. Based on this information, the user decides on an action and provides feedback to the system as needed. Specific actions include interpreting the results on the screen and considering countermeasures. The output of this step is the user's feedback and actions.

[0712] (Application Example 2)

[0713] 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".

[0714] Providing appropriate responses and suggestions that address residents' emotional states within the home environment is believed to contribute to reducing stress and improving their quality of life. However, conventional home systems have struggled to accurately recognize emotional states in real time and optimize interactions based on that information. This has resulted in problems in responding appropriately to residents' needs.

[0715] 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.

[0716] In this invention, the server includes means for preprocessing user behavior data and voice data acquired from a data collection device; means for generating a machine learning model that learns using the preprocessed data; means for analyzing the data in real time using the retrained model to identify the user's emotional state; and means for determining a response corresponding to the identified emotional state and notifying the user. This makes it possible to provide appropriate responses in real time that are in line with the user's emotions in a home environment.

[0717] A "data acquisition device" is a device consisting of hardware and software for acquiring user behavior data and voice data.

[0718] "Preprocessing" refers to the process of applying noise reduction, standardization, and other treatments to collected data to prepare it for analysis.

[0719] A "machine learning model" refers to an algorithm that uses pre-processed data to learn and recognize or predict specific patterns.

[0720] "Retraining" is the process of performing additional training to improve an existing machine learning model based on evaluation results.

[0721] "Real-time analysis" refers to the immediate processing of collected data to instantly identify the user's emotional state and other relevant information.

[0722] "Emotional state" refers to the psychological or emotional condition identified from the user's actions and tone of voice.

[0723] "Notification" refers to the act of communicating analysis results and response content to the user.

[0724] The system designed to realize this application is designed to understand the user's emotional state in real time within the home and provide appropriate responses and suggestions based on that understanding.

[0725] The server first acquires user behavior and voice data via a data collection device. This data is sent to the server in real time and undergoes preprocessing such as noise reduction and standardization. The preprocessed data is then input into a generative AI model for emotion recognition using machine learning frameworks such as TensorFlow or PyTorch. The model identifies the user's emotional state from the data and classifies it into categories such as positive, negative, and neutral.

[0726] The device makes appropriate suggestions to the user based on emotional state information received from the server. For example, if the device identifies that the user is stressed, it can respond by suggesting that the user play relaxing music.

[0727] Users view the analysis results through a user interface on their device and provide feedback on the suggestions offered. This feedback is sent to the server and used as part of the model's retraining process to improve the overall system accuracy.

[0728] For example, if a user comes home tired after a long day at work, the system can analyze the user's tone of voice and facial expression and suggest, "Shall I play some music to help you relax?"

[0729] An example of a prompt message is: "Design a robot application that offers stress-reducing suggestions within the home based on the user's emotions. For example, it should include a feature that suggests relaxation music when the user is feeling stressed."

[0730] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0731] Step 1:

[0732] The server acquires user motion data and voice data in real time from data acquisition devices. Since the data may be raw and contain noise, preprocessing is performed immediately after data acquisition. The server receives data streams from cameras and microphones as input and generates standardized, pre-processed data as output, with noise removed.

[0733] Step 2:

[0734] The server inputs preprocessed data into a generative AI model for emotion recognition. Here, TensorFlow or PyTorch is used to identify specific emotion patterns from the data. Using preprocessed data as input, the output is a label indicating the user's emotional state (positive, negative, neutral, etc.). Algorithms such as multilayer neural networks are used in this process.

[0735] Step 3:

[0736] The server determines what suggestions or responses to offer the user based on the identified emotional state. The input is the recognized emotional label, and the output is the corresponding response or suggestion. At this stage, for example, if the user is feeling stressed, an action such as suggesting playing relaxation music might be selected.

[0737] Step 4:

[0738] The terminal notifies the user of the suggestions received from the server. A user interface is utilized here, conveying the suggestions or responses to the user visually or audibly. The input is the suggestions from the server, and the output is the user receiving visualized or audible information.

[0739] Step 5:

[0740] Users provide feedback on the suggestions offered through the terminal's user interface. The input is user feedback information, and the output is data sent to the server. This feedback is used in subsequent relearning processes.

[0741] Step 6:

[0742] The server retrains the generative AI model based on accumulated feedback data to improve the system's accuracy. The input is feedback data and past model states, and the output is the updated generative AI model. Retraining is performed periodically to improve the model's performance.

[0743] 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.

[0744] 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.

[0745] 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.

[0746] 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.

[0747] 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.

[0748] 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.

[0749] 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.

[0750] 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.

[0751] 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."

[0752] 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.

[0753] 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.

[0754] 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.

[0755] 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.

[0756] 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.

[0757] 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.

[0758] 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.

[0759] 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.

[0760] 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.

[0761] 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.

[0762] 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.

[0763] 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 to be incorporated by reference.

[0764] The following is further disclosed regarding the embodiments described above.

[0765] (Claim 1)

[0766] A means for preprocessing data acquired from a data acquisition device,

[0767] A means for generating a machine learning model that performs training using preprocessed data,

[0768] A means for evaluating the generated machine learning model and retraining the model based on the evaluation results,

[0769] A method for analyzing data in real time using a retrained model,

[0770] A means of notifying the user of the analysis results,

[0771] A system that includes this.

[0772] (Claim 2)

[0773] The system according to claim 1, which performs real-time processing that utilizes data acquired from a data acquisition device for anomaly detection.

[0774] (Claim 3)

[0775] The system according to claim 1, comprising a user interface means for providing information including analysis results to the user.

[0776] "Example 1"

[0777] (Claim 1)

[0778] Means for preprocessing information acquired from an information acquisition device,

[0779] A means for generating a machine learning model that performs learning using preprocessed information,

[0780] A means for evaluating the generated machine learning model and retraining the model based on the evaluation results,

[0781] A method for performing format conversion while simultaneously removing noise,

[0782] A means of analyzing information in real time using a retrained model,

[0783] A means of notifying a human of the analysis results via a terminal,

[0784] A system that includes this.

[0785] (Claim 2)

[0786] The system according to claim 1, which performs real-time processing that utilizes information acquired from an information acquisition device for anomaly detection.

[0787] (Claim 3)

[0788] The system according to claim 1, comprising a display interface means for providing information including analysis results to a human.

[0789] "Application Example 1"

[0790] (Claim 1)

[0791] A means for preprocessing data acquired from a data acquisition device,

[0792] A means for generating a machine learning model that performs training using preprocessed data,

[0793] A means for evaluating the generated machine learning model and retraining the model based on the evaluation results,

[0794] A method for analyzing data in real time using a retrained model,

[0795] A means of notifying the user of the analysis results via a user interface,

[0796] A means for transmitting event information generated based on the analysis results to a remote device using data communication technology,

[0797] A system that includes this.

[0798] (Claim 2)

[0799] The system according to claim 1, which performs real-time processing to use data acquired from a data acquisition device for anomaly detection and notifies a remote device of the detected anomaly information.

[0800] (Claim 3)

[0801] The system according to claim 1, comprising a user interface means for providing information including analysis results to the user, and supporting interaction for the user to quickly perform operations based on the analysis results.

[0802] "Example 2 of combining an emotion engine"

[0803] (Claim 1)

[0804] A means for preprocessing information collected from data acquisition equipment,

[0805] A means for generating an artificial intelligence model that learns using preprocessed information,

[0806] A means for evaluating the generated artificial intelligence model and retraining the model based on the evaluation results,

[0807] A means of analyzing information in real time using a retrained model,

[0808] A means for notifying the user of the analysis results and emotional state, and for optimizing the presentation method,

[0809] A system that includes this.

[0810] (Claim 2)

[0811] The system according to claim 1, which performs real-time processing that utilizes information collected from data acquisition equipment for anomaly detection.

[0812] (Claim 3)

[0813] The system according to claim 1, comprising a user terminal means that provides the user with information including analysis results and emotional data.

[0814] "Application example 2 when combining with an emotional engine"

[0815] (Claim 1)

[0816] Means for preprocessing user action data and voice data acquired from a data acquisition device,

[0817] A means for generating a machine learning model that performs training using preprocessed data,

[0818] A means for evaluating the generated machine learning model and retraining the model based on the evaluation results,

[0819] A means of analyzing data in real time using a retrained model to identify the user's emotional state,

[0820] A means for determining a response according to the identified emotional state and notifying the user,

[0821] A system that includes this.

[0822] (Claim 2)

[0823] The system according to claim 1, which performs real-time processing that utilizes data acquired from a data acquisition device for anomaly detection and emotion identification.

[0824] (Claim 3)

[0825] The system according to claim 1, comprising a user interface means for providing analysis results and emotional state information to the user. [Explanation of symbols]

[0826] 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. A means for preprocessing data acquired from a data acquisition device, A means for generating a machine learning model that performs training using preprocessed data, A means for evaluating the generated machine learning model and retraining the model based on the evaluation results, A method for analyzing data in real time using a retrained model, A means of notifying the user of the analysis results, A system that includes this.

2. The system according to claim 1, which performs real-time processing that utilizes data acquired from a data acquisition device for anomaly detection.

3. The system according to claim 1, comprising a user interface means for providing information including analysis results to the user.

Images