System

Abstract

A system is provided.SOLUTION: A system, comprising: means for detecting an emotion from a user's voice using an emotion detection model; means for generating appropriate feedback based on the detected emotion using a mentoring generation model; and means for audibly conveying the generated feedback to the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In modern society, many people experience high stress levels and emotional ups and downs. This increases mental health problems and negatively impacts their quality of life and productivity. In addition, it is often difficult for people to discuss their emotional issues with others, limiting opportunities to receive appropriate support. Given this background, there is a need for a system that can quickly detect users' emotions and provide appropriate feedback. [Means for solving the problem]

[0005] The present invention provides a system including a means for detecting emotions from a user's voice data using an emotion detection model, a means for generating appropriate feedback based on the detected emotions using a mentoring generation model, and a means for communicating the generated feedback to the user by voice. The emotion detection model detects emotions by analyzing the tone, speed, and volume of the voice, and the mentoring generation model generates specific advice based on the detected emotions. The system further includes a means for receiving the generated feedback in text format and converting it into voice using a speech synthesis engine, and also includes a communication means for performing emotion detection and feedback in real time between the user and the system. This allows users to quickly detect emotional issues and receive timely and appropriate support, thereby improving their mental health.

[0006] An "emotion detection model" is an algorithm or machine learning model for identifying a user's emotions from voice data.

[0007] A "mentoring generative model" is an algorithm or machine learning model that generates appropriate feedback and advice based on the emotions identified by the emotion detection model.

[0008] "Voice data" refers to data obtained by digitizing the user's voice.

[0009] "Feedback" refers to output data that provides useful information or advice regarding the user's emotions.

[0010] "Speech synthesis engine" means a system or technology for converting textual feedback into speech.

[0011] "Communication means" refers to the technology or protocol used to send and receive data between a user and a server, or between components.

[0012] An "emotion label" is a tag or name that indicates the type of emotion identified by an emotion detection model.

[0013] A "user" is an individual who uses the system to receive emotion detection and feedback.

[0014] "Real time" refers to the timing when data is acquired, processed, and output without delay. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12]FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0016] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0017] First, the terms used in the following description will be explained.

[0018] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0019] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0020] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0021] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0022] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0023] [First embodiment]

[0024] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0025] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0026] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0027] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0028] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0029] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0030] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0031] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0033] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0034] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0035] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0036] As an embodiment of the present invention, a system using an emotion detection model and a mentoring generation model will be described. This system analyzes a user's voice data in real time and provides appropriate feedback.

[0037] Overall system overview

[0038] The system consists of three main components:

[0039] 1. Emotion detection model

[0040] 2. Mentoring Generative Model

[0041] 3. Speech synthesis engine

[0042] Emotion detection model

[0043] The emotion detection model identifies emotions from a user's voice data. When a user speaks to a device (such as a smart speaker), the voice data is captured and sent to a server. The server inputs this voice data into the emotion detection model and generates an emotion label. The emotion detection model analyzes features such as tone, speed, and volume of the voice to identify the emotion the user is feeling (e.g., anger, sadness, joy, etc.).

[0044] Mentoring Generative Model

[0045] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. The mentoring generative model generates useful feedback for the user according to the identified emotion. For example, if the user is feeling stressed, it generates advice such as "Take a deep breath and relax."

[0046] Text-to-speech engine

[0047] The generated feedback is returned in text format to the server, which then inputs this text into a speech synthesis engine to convert it into a speech format that is easy for the user to understand. The converted speech is then transmitted to the user via the terminal.

[0048] Specific use cases

[0049] 1. Emotion detection and feedback

[0050] 1. The user speaks into the device.

[0051] For example: "I'm really tired today."

[0052] 2. The device captures this audio and sends it to the server as audio data.

[0053] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[0054] 4. The server inputs the generated emotion labels into the mentoring generative model.

[0055] 5. The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as the message "Please take an early rest today."

[0056] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[0057] 7. The device plays the converted audio to the user.

[0058] For example: "Please go to bed early today."

[0059] 8. Users can take appropriate action based on the feedback.

[0060] In this way, users can use the system to quickly detect their emotions and receive appropriate advice and feedback at that moment, which is expected to serve as a useful tool for maintaining mental health.

[0061] The processing flow will be explained below.

[0062] Step 1:

[0063] The user speaks to the device, for example, saying, "I'm really tired today."

[0064] Step 2:

[0065] The device captures this audio and converts it into digital audio data.

[0066] Step 3:

[0067] The terminal transmits digital audio data to the server.

[0068] Step 4:

[0069] The server inputs the received audio data into the emotion detection model.

[0070] Step 5:

[0071] An emotion detection model analyzes the tone, rate, and volume of the speech and generates an emotion label (e.g., "fatigue").

[0072] Step 6:

[0073] The server inputs the generated emotion labels into the mentoring generative model.

[0074] Step 7:

[0075] Based on the emotion labels, the mentoring generative model generates appropriate feedback such as "Please take a rest early today."

[0076] Step 8:

[0077] The server inputs the generated message into a speech synthesis engine and converts it into speech format.

[0078] Step 9:

[0079] The terminal plays back the converted audio data and provides feedback to the user.

[0080] Step 10:

[0081] The user listens to the audio feedback from the device and takes appropriate action, such as taking an early rest.

[0082] Through this series of steps, users' emotions are detected in real time and they can receive appropriate advice on the spot.

[0083] Example 1

[0084] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0085] Conventional emotion detection systems have the drawback of being unable to provide real-time feedback and promptly provide appropriate advice based on the user's emotional state. Another problem is that multiple methods for improving the accuracy of emotion detection and the quality of feedback are not integrated.

[0086] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0087] In this invention, the server includes means for acquiring and transmitting user voice data to the server, means for the server to receive the voice data, input the voice data to an emotion detection model, and analyze the voice features to generate an emotion label, means for the server to input the emotion label to a mentoring generation model and generate an appropriate feedback message, means for inputting the generated feedback message to a voice synthesis engine and converting it into a voice format, and means for transmitting the converted voice to the terminal and playing it back to the user, thereby enabling real-time emotion detection and immediate feedback provision.

[0088] A "user" is a human entity that provides speech data to the system and receives feedback.

[0089] A "terminal" is a device for capturing and transmitting audio data to a server, and is a device for playing back the feedback audio.

[0090] A "server" is a computing system that receives audio data, processes the data using the emotion detection model and the mentoring generation model, and generates feedback.

[0091] An "emotion detection model" is an artificial intelligence model that analyzes characteristics of voice data, such as tone, speed, and volume, to identify a user's emotions.

[0092] An "emotion label" is data that indicates the emotional state of a user extracted from voice data by an emotion detection model.

[0093] The "Mentoring Generative Model" is an artificial intelligence model that generates appropriate feedback messages based on emotion labels.

[0094] A "feedback message" is a text message containing advice or instructions that correspond to the user's emotional state.

[0095] "Speech synthesis engine" means software or a device that converts textual feedback messages into speech.

[0096] "Converted speech" is speech data generated from a text message by a speech synthesis engine.

[0097] "Real-time" refers to processing or feedback that occurs immediately or with very little delay.

[0098] This invention is a system that uses an emotion detection model, a mentoring generation model, and a speech synthesis engine to analyze a user's voice data in real time and provide appropriate feedback. A specific embodiment of this system is described below.

[0099] Overall system configuration

[0100] The system consists of the following main components:

[0101] 1. Emotion detection model

[0102] 2. Mentoring Generative Model

[0103] 3. Speech synthesis engine

[0104] Emotion detection model

[0105] The user provides voice data to the device. For example, the user might say, "I'm really tired today." The device captures the voice data and sends it to the server. The server inputs the received voice data into an emotion detection model. The emotion detection model analyzes features such as tone, speed, and volume of the voice and generates an emotion label. The specific software used is, for example, LibROSA, which extracts the features of the voice data, and a generative AI model based on TENSORFLOW (registered trademark) or PyTorch.

[0106] Mentoring Generative Model

[0107] The server obtains the emotion labels obtained from the emotion detection model and inputs them into the mentoring generation model. The mentoring generation model generates useful feedback for the user based on the identified emotion. For example, if the user is feeling stressed, the mentoring generation model generates a message such as "Please take an early rest today." The generative AI model used here is GPT-3 (registered trademark).

[0108] Text-to-speech engine

[0109] The generated feedback message is returned in text format to the server. The server then inputs this text into a speech synthesis engine, which converts it into a speech format that is easy for the user to understand. Specific software used includes Google® Text-to-Speech API and Amazon Polly. The converted speech is sent to the device and played back to the user.

[0110] Specific use cases

[0111] A specific example of the system's operation is shown below.

[0112] 1. The user speaks to the device, saying, "I'm really tired today."

[0113] 2. The device captures this audio and sends it to the server as audio data.

[0114] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[0115] 4. The server inputs the generated emotion labels into the mentoring generative model.

[0116] 5. The mentoring generative model generates an appropriate feedback message based on the emotion label "fatigue," such as "Please take an early rest today."

[0117] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[0118] 7. The device plays the converted audio to the user, telling them, "Please go to bed early today."

[0119] 8. Users can take appropriate action based on the feedback.

[0120] A concrete example of a prompt for a generative AI model

[0121] 1. Example prompt for emotion detection model:

[0122] "Based on this audio data, identify the user's current emotion."

[0123] 2. Example prompts for the Mentoring Generative Model:

[0124] "When the emotion label is 'fatigue', generate an appropriate feedback message."

[0125] 3. Example prompt for the speech synthesis engine:

[0126] "Please convert the text 'Please go to bed early today' into audio format."

[0127] In this way, users receive rapid, real-time feedback from the system and can take action to improve their physical and mental health.

[0128] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0129] Step 1:

[0130] The user provides voice data to the terminal, which then captures the voice with a microphone and converts it into digital data.

[0131] Input: User's voice

[0132] Output: Digitized audio data

[0133] Specific operations: Records audio with a microphone, converts the audio data into digital format, and stores the digital data in temporary memory.

[0134] Step 2:

[0135] The device prepares the captured audio data to be sent to the server, dividing it into packets and sending them via TCP / IP.

[0136] Input: Digitized voice data

[0137] Output: Audio data packets to the server

[0138] Specific operation: Packetizes digital data into a specific format and sends it over a network to a server.

[0139] Step 3:

[0140] The server analyzes the received audio data packets and reconstructs the audio data, which is then input into the emotion detection model.

[0141] Input: Voice data packets

[0142] Output: Reconstructed speech data, input data for emotion detection model

[0143] Specific operation: Receive packets, reassemble them appropriately, and combine them into a single audio file. Input the audio into the emotion detection model.

[0144] Step 4:

[0145] The server inputs the audio data into an emotion detection model, analyzes audio features (e.g., tone, rate, volume) and generates emotion labels.

[0146] Input: Reconstructed audio data

[0147] Output: Emotion label

[0148] Specific operation: Extract features from audio data using a library such as LibROSA. Input the extracted features into a generative AI model (e.g., TensorFlow or PyTorch) to obtain emotion labels.

[0149] Step 5:

[0150] The server inputs the emotion labels into a mentoring generative model to generate appropriate feedback messages.

[0151] Input: Emotion label

[0152] Output: Feedback message

[0153] Specific action: The emotion label is input as a prompt sentence into the mentoring generative model to obtain a feedback message (e.g., "Please go to bed early today").

[0154] Step 6:

[0155] The server inputs the generated feedback message into a speech synthesis engine and converts it into speech format.

[0156] Input: Feedback message

[0157] Output: Audio data

[0158] Specific behavior: Converts text messages into speech using the Google Text-to-Speech API, Amazon Polly, etc.

[0159] Step 7:

[0160] The server transmits the feedback message converted into audio format to the terminal.

[0161] Input: Audio data

[0162] Output: Audio data packets

[0163] Specific operation: The generated voice data is packetized for transmission to the terminal and sent via TCP / IP.

[0164] Step 8:

[0165] The device analyzes and reconstructs the received audio data and plays it back through the speaker, thereby providing feedback to the user.

[0166] Input: Voice data packets

[0167] Output: Reconstructed speech data, speech feedback to the user

[0168] Specific operation: The received packets are reconstructed and compiled into a single audio file, which is then played back through the speaker to the user.

[0169] (Application example 1)

[0170] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0171] Conventional food delivery systems have difficulty providing appropriate support and meal suggestions based on the user's emotional state, and have been unable to increase the user's psychological satisfaction. Furthermore, when a user feels fatigued or stressed, there is a lack of a way to provide real-time advice tailored to that emotional state. Therefore, there is a need for improved user experience and psychological health support.

[0172] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0173] In this invention, the server includes means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for communicating the generated feedback to the user by voice, means for including meal suggestions as part of the feedback, and means for including an interface for the user to order food. This makes it possible to provide appropriate feedback and advice according to the user's emotional state, and further to suggest meals and order food delivery according to the emotions.

[0174] An "emotion detection model" is an artificial intelligence model that analyzes a user's voice data and identifies emotions from characteristics such as tone, speed, and volume of the voice.

[0175] The "mentoring generation model" is an artificial intelligence model that generates useful feedback and advice for users based on detected emotions.

[0176] "Speech synthesis engine" refers to software that converts the generated textual feedback into speech.

[0177] "Feedback" refers to appropriate advice or guidance provided based on the user's emotional state.

[0178] "Meal suggestion" refers to selecting and suggesting meals suitable for the user based on the user's emotional state.

[0179] An "interface" refers to the operating screen and input means that a user uses to interact with the system and order food.

[0180] "Voice data" refers to the spoken words or sounds of a user and is digital data captured by a voice recognition system.

[0181] An "emotion label" is an indicator or tag that indicates a user's emotion identified by an emotion detection model.

[0182] As an embodiment of the present invention, a system for analyzing user voice data, detecting emotions, and providing appropriate feedback based on the emotions in a food delivery system will be described. This system mainly consists of the following components:

[0183] 1. Emotion detection model

[0184] 2. Mentoring Generative Model

[0185] 3. Speech synthesis engine

[0186] 4. Meal suggestion tools

[0187] 5. Food ordering interface

[0188] Emotion detection model

[0189] The server receives voice data from users speaking into devices such as smartphones. This voice data is input into an emotion detection model, which analyzes features such as tone, speed, and volume of the voice and generates emotion labels, such as "fatigue" or "joy."

[0190] Mentoring Generative Model

[0191] The server generates appropriate feedback using a mentoring generation model based on the emotion labels obtained from the emotion detection model. This feedback includes advice and guidance based on the emotion the user is feeling. For example, if the user has an emotion label of "fatigue," feedback including a specific meal suggestion such as "hot soup would be good" is generated.

[0192] Text-to-speech engine

[0193] The generated feedback is sent from the server to a speech synthesis engine, which converts the feedback text into speech, which is then transmitted to the user via the terminal.

[0194] Meal suggestion tool

[0195] The feedback generated by the mentoring generative model includes specific meal suggestions based on the user's emotional state, which are structured to help the user choose appropriate meals.

[0196] Food Ordering Interface

[0197] An interface is provided for the user to order food based on the suggested meal menu, and the user can easily use this interface to order the suggested meal.

[0198] Specific examples

[0199] For example, suppose a user says to the device, "I'm tired today, so I want to eat something comforting." The device captures the voice data and sends it to the emotion detection model. The server obtains the emotion label "fatigue" and uses the mentoring generative model to generate feedback such as "If you're feeling tired, a warm soup would be good." This feedback is sent to the speech synthesis engine, converted into audio format, and transmitted to the user. The user can easily order the suggested meal through the provided food ordering interface.

[0200] Prompt Sentence Examples

[0201] "User voice text: 'I'm tired today and I want to eat something comforting.'

[0202] Emotion detection model output: 'Fatigue'

[0203] Generate mentoring messages based on this emotion.”

[0204] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0205] Step 1:

[0206] The user speaks to the device and inputs voice data. For example, if the user says, "I'm tired today, so I want to eat something soothing," the voice is captured by the device.

[0207] Step 2:

[0208] The device sends the captured audio data to the server, which then becomes the input data for the system.

[0209] Step 3:

[0210] The server inputs the received voice data into an emotion detection model, which analyzes the tone, speed, volume, etc. of the voice to identify the user's emotion label. In this step, an emotion label such as "fatigue" is output.

[0211] Step 4:

[0212] The server inputs the generated emotion labels into the mentoring generative model, which then generates an appropriate feedback message based on the emotion labels. For example, for the emotion label "fatigue," the model generates the message "If you're feeling tired, warm soup would be good."

[0213] Step 5:

[0214] The server inputs the generated feedback message into a speech synthesis engine, which converts the text message into a voice message. Specifically, the text "If you're feeling tired, a warm soup would be good" is converted into voice data.

[0215] Step 6:

[0216] The server transmits the converted voice data to the terminal, which then plays the voice data and transmits it to the user, so that the user receives voice feedback.

[0217] Step 7:

[0218] The server generates a message containing specific meal suggestions as part of the feedback, such as a specific menu item like "hot soup," in a format that makes it easy for the user to choose.

[0219] Step 8:

[0220] The terminal provides a user interface that displays the suggested meal menu and provides a means for the user to perform operations to order it. The user can use this interface to easily order the suggested meal.

[0221] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0222] As an embodiment of the present invention, we will explain a system that combines an emotion engine that recognizes a user's emotions using information such as facial expressions and body movements in addition to voice data. This system quickly detects a user's diverse emotional information and provides appropriate feedback in real time.

[0223] Overall system overview

[0224] The system consists of four main components:

[0225] 1. Emotion Engine

[0226] 2. Emotion Detection Model

[0227] 3. Mentoring Generative Model

[0228] 4. Speech synthesis engine

[0229] Emotion Engine

[0230] The emotion engine recognizes emotions from the user's voice data, facial expressions, and body movements. When the user is close to the device, the device's camera and sensors capture facial expressions and body movements. In addition, voice data is also acquired. This data is sent to the server and input into the emotion engine. The emotion engine integrates this multidimensional data and recognizes the user's emotions (e.g., joy, anger, sadness, surprise, etc.) with high accuracy.

[0231] Emotion detection model

[0232] The emotion detection model further analyzes the voice data based on the emotional information provided by the emotion engine. This model further analyzes voice characteristics such as tone, speed, and volume to determine the emotion. It can also refer to the user's past emotional history to predict the user's current emotional state.

[0233] Mentoring Generative Model

[0234] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. This model generates advice according to the identified emotion and provides useful feedback to the user.

[0235] Text-to-speech engine

[0236] The generated feedback is returned to the server in text format and converted into audio format by inputting it into a speech synthesis engine. The device plays back this audio data and conveys the feedback to the user.

[0237] Specific use cases

[0238] 1. Recognizing emotions and providing feedback

[0239] 1. As the user speaks to the device, data including facial expressions and body movements is captured by the device's camera and sensors.

[0240] For example: "I'm really tired today."

[0241] 2. The device transmits this voice data, facial expression data, and body movement data to the server.

[0242] 3. The server inputs the received multidimensional data into the emotion engine, which analyzes the data and generates an emotion label, such as "fatigue."

[0243] 4. The server re-inputs the audio data, along with the emotion labels obtained from the emotion engine, into the emotion detection model, which analyzes the audio features and reconfirms the results from the emotion engine.

[0244] 5. The server requests the mentoring model to generate feedback based on the determined emotion label. In response to the emotion "fatigue," the mentoring model generates advice such as "Take an early rest today."

[0245] 6. The server inputs this advice into a speech synthesis engine and converts it into voice data.

[0246] 7. The device plays the generated audio data and provides feedback to the user.

[0247] For example: "Please go to bed early today."

[0248] 8. The user follows the feedback and takes appropriate action.

[0249] This system allows users to receive accurate feedback in real time based on a variety of emotional data, which is expected to provide more effective support for users' mental health.

[0250] The processing flow will be explained below.

[0251] Step 1:

[0252] As a user speaks to the device, the device's camera and sensors capture the user's facial expressions and body movements. For example, the user might say, "I'm really tired today."

[0253] Step 2:

[0254] The device digitizes the captured voice data, facial expression data, and body movement data and transmits them to a server.

[0255] Step 3:

[0256] The server inputs the received voice data, facial expression data, and body movement data into the emotion engine, which integrates and analyzes the data to identify the user's emotional state.

[0257] Step 4:

[0258] Based on the analysis results, the emotion engine generates an emotion label called "fatigue" and returns it to the server.

[0259] Step 5:

[0260] The server inputs the emotion labels and audio data obtained from the emotion engine into the emotion detection model, which analyzes the tone, rate, and volume of the audio and double-checks the results from the emotion engine.

[0261] Step 6:

[0262] The emotion detection model determines the emotion "fatigue" and returns the emotion label to the server.

[0263] Step 7:

[0264] The server inputs the determined emotion labels into the mentoring generative model.

[0265] Step 8:

[0266] The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as "Please take an early rest today."

[0267] Step 9:

[0268] The server inputs the generated feedback into a speech synthesis engine, which converts the text-format feedback into speech data.

[0269] Step 10:

[0270] The terminal plays back the converted audio data and provides feedback to the user.

[0271] Step 11:

[0272] The user listens to the audio feedback from the device and takes appropriate action based on it, such as taking an early rest.

[0273] Through this series of steps, users can have their emotions detected from multiple angles using the emotion engine, and receive optimal feedback in real time.

[0274] Example 2

[0275] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0276] Conventional emotion recognition systems detect emotions based solely on the user's voice data, resulting in insufficient emotion recognition accuracy. Furthermore, the quality and applicability of the feedback provided is limited, resulting in insufficient user response. Therefore, there is a need for a system that can improve emotion recognition accuracy and provide more appropriate feedback.

[0277] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0278] In this invention, the server includes means for recognizing emotions from multiple data sources including facial expressions and body movements, means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, and means for transmitting the generated feedback to the user by voice, thereby improving the accuracy of emotion recognition and enabling more appropriate and useful feedback to be provided to the user in real time.

[0279] - "Multiple data sources including facial expressions and body movements" is a general term for input methods such as sensors, cameras, and microphones that collect multiple types of data, such as a user's facial expressions, body movements, and even voice data.

[0280] "Means for recognizing emotions" refers to a system or algorithm that analyzes collected voice data, facial expression data, and body movement data to identify a user's emotions.

[0281] An "emotion detection model" refers to a machine learning model or algorithm that analyzes a user's voice data and detects emotions from its characteristics, such as tone, speed, and volume.

[0282] "Mentoring generative model" means a system or algorithm that generates appropriate feedback and advice for a user based on emotional information obtained from an emotion detection model.

[0283] "Means for generating feedback" refers to software or a program that uses an emotion detection model or a mentoring generation model to generate advice or guidance according to the user's emotions.

[0284] "Means for transmitting the generated feedback to the user by voice" refers to a system that uses a speech synthesis engine to convert text-based feedback into audio format and transmits it to the user through speakers or earphones.

[0285] An embodiment of the present invention is a system that recognizes emotions using a user's voice data, facial expressions, and body movements, and provides appropriate feedback in real time.

[0286] System Overview

[0287] The system mainly consists of the following elements:

[0288] 1. Data sources including facial expressions and body movements

[0289] 2. Means of recognizing emotions (emotion engine)

[0290] 3. Emotion Detection Model

[0291] 4. Mentoring Generative Model

[0292] 5. Speech synthesis engine

[0293] System configuration

[0294] 1. Data Source:

[0295] The user speaks to the device, and the device's built-in camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data.

[0296] As a specific example, consider the case where a user says, "I'm really tired today." At this time, the device captures the user's tired facial expression, slumped shoulders, and other aspects of their posture, and the microphone records their voice.

[0297] 2. Means of Emotion Recognition (Emotion Engine):

[0298] The device sends the collected data to a server in real time. The emotion engine on the server analyzes this multidimensional data and recognizes the user's emotions. Specifically, it uses facial expression recognition algorithms and motion analysis algorithms.

[0299] 3. Emotion detection model:

[0300] The server inputs the voice data into an emotion detection model, which analyzes voice characteristics such as tone, speed, and volume to detect emotions. It also uses past emotion data to improve the accuracy of emotion detection.

[0301] 4. Mentoring generative model:

[0302] The server receives the output of the emotion detection model and generates appropriate feedback using the mentoring generation model. The generated feedback includes advice and comments according to the user's emotional state. For example, the generated feedback might be "Please go to bed early today."

[0303] 5. Speech synthesis engine:

[0304] The server inputs the generated feedback into a speech synthesis engine, converting the text into natural-sounding speech data, which is then sent to the device.

[0305] 6. Providing feedback:

[0306] The device plays back the voice data sent from the speech synthesis engine and provides feedback to the user. For example, a voice message saying "Please go to bed early today" is played through the speaker or earphones.

[0307] Prompt Sentence Examples

[0308] Below are some example prompts to ensure the system works properly:

[0309] Capture the voice data, facial expressions, and body movements of a user saying, "I'm really tired today," and input them into the emotion engine. After that, once the emotion label "fatigue" is generated, reanalyze it with the emotion detection model, and finally generate the advice "Please take an early rest today," and convert it into voice data with the speech synthesis engine.

[0310] This system recognizes the user's emotions with high accuracy from multiple data sources, including not only their voice but also their facial expressions and body movements, and provides real-time feedback according to their current state, making it possible to more appropriately support the user's physical and mental state.

[0311] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0312] Step 1:

[0313] The user speaks to the device, and the device's camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data (input: user's voice, facial expressions, and body movements; output: multiple data sources).

[0314] Step 2:

[0315] The device transmits the captured facial expression data, body movement data, and voice data to the server in real time (input: multiple data sources, output: data transfer to the server).

[0316] Step 3:

[0317] The server inputs the received multidimensional data into the emotion engine, which then integrates the data and identifies the user's emotion using facial expression recognition and motion analysis algorithms (input: data transferred to the server, output: recognized emotion label).

[0318] Step 4:

[0319] The server inputs the emotion label and audio data output from the emotion engine into the emotion detection model, which analyzes features such as tone, speed, and volume of the audio to re-detect the emotion in detail (input: emotion label, audio data, output: confirmed emotion label).

[0320] Step 5:

[0321] The server inputs the determined emotion label into the mentoring generative model and generates feedback. The mentoring generative model generates appropriate advice based on the identified emotion (input: determined emotion label, output: generated feedback).

[0322] Step 6:

[0323] The server inputs the generated feedback into a speech synthesis engine and converts the text feedback into speech data (input: generated feedback text, output: speech data).

[0324] Step 7:

[0325] The device receives the voice data sent from the server and provides feedback to the user through a speaker or earphones (input: voice data, output: voice message to the user).

[0326] Specific operation example

[0327] 1. The user says to the device, "I'm really tired today."

[0328] 2. The device collects the user's facial expressions, body movements, and voice data and sends it to the server.

[0329] 3. The server uses an emotion engine to recognize the emotion "fatigue."

[0330] 4. The server performs detailed analysis of the voice data using an emotion detection model to reconfirm the emotion.

[0331] 5. The server uses a mentoring generation model to generate advice such as "Take a rest early today," and converts it into voice data using a speech synthesis engine.

[0332] 6. The device plays the generated voice data and provides the user with feedback such as "Please go to bed early today."

[0333] (Application example 2)

[0334] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0335] In autonomous vehicles, the challenge is to monitor the emotional state of passengers in real time and improve their comfort and safety. This will provide an environment where passengers can use autonomous vehicles safely and without anxiety or stress, even for long periods of time.

[0336] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0337] In this invention, the server includes means for detecting emotions from the user's voice data, facial expression data, and body movement data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for providing the feedback to the infotainment system of the autonomous vehicle, and means for communicating the generated feedback to the user by voice. This makes it possible to recognize the passenger's emotional state in real time and provide advice or adjust environmental settings (music, lighting, interior temperature) according to the state.

[0338] An "emotion detection model" is a system for detecting emotions by analyzing a user's voice data, facial expression data, and body movement data.

[0339] The "mentoring generation model" is a system that generates appropriate feedback and advice based on emotions detected by the emotion detection model.

[0340] An "infotainment system" is a system that provides information and entertainment functions and is installed in an autonomous vehicle.

[0341] "Feedback" refers to advice and information generated by the emotion detection model and mentoring generation model according to the user's emotional state.

[0342] "Voice data" is a digital representation of a voice signal containing the content of a user's speech.

[0343] "Facial expression data" refers to a facial expression of a user captured by a photographing device such as a camera and expressed in digital form.

[0344] "Body movement data" refers to the user's body movements captured by sensors or other devices and expressed in digital form.

[0345] An "automated vehicle" is a vehicle whose driving operation is performed by an automatic control system.

[0346] This paper describes an embodiment of the present invention. The system analyzes a user's voice data, facial expression data, and body movements to recognize emotions and improve passenger comfort and safety in an autonomous vehicle. Specifically, it consists of the following four main components:

[0347] 1. Emotion Engine

[0348] 2. Emotion Detection Model

[0349] 3. Mentoring Generative Model

[0350] 4. Speech synthesis engine

[0351] Emotion Engine

[0352] The emotion engine captures the user's voice data, facial expression data, and body movement data. When a user gets into an autonomous vehicle, cameras and sensors installed in the vehicle capture this data and send it to a server. This uses tools such as PyAudio to capture voice data and OpenCV to capture facial expressions and body movements. The data sent to the server is analyzed as multidimensional data by the emotion engine to recognize the user's emotions.

[0353] Emotion detection model

[0354] The emotion detection model uses the emotional information provided by the emotion engine to perform detailed analysis of voice characteristics such as tone, speed, volume, facial expressions, and body movements. This allows it to accurately determine the user's emotional state. The model is built using machine learning frameworks such as TensorFlow and Keras. For example, if the user is relaxed, it generates a label such as "calm," and if they are nervous, it generates a label such as "nervous."

[0355] Mentoring Generative Model

[0356] The mentoring generative model generates appropriate feedback based on the emotion labels obtained from the emotion detection model. The server inputs the emotion labels received from the emotion detection model into a generative AI model (e.g., GPT-2) and generates feedback in text format. This feedback may include advice based on the passenger's emotions or adjustments to environmental settings (music, lighting, interior temperature, etc.).

[0357] Text-to-speech engine

[0358] The generated feedback is input into a speech synthesis engine and converted into voice format. The server then transmits this voice data to the passenger through the infotainment system in the autonomous vehicle. Specific examples of feedback could be, "Take a short break today" or "We'll lower the temperature inside the car a little."

[0359] Specific use cases

[0360] For example, if a user in an autonomous vehicle says, "Oh, I'm really tired today," a camera and microphone installed in the infotainment system capture their voice, facial expressions, and body movements. The voice and facial expression data are sent to a server and analyzed by an emotion engine. If the emotion detection model determines that the user is "fatigue," the mentoring generation model generates feedback such as "Please take an early rest today," which is then converted into voice data by a speech synthesis engine and provided to the user. This feedback contributes to creating an environment where the user can feel safe and relaxed.

[0361] Example prompt sentence:

[0362] "The passenger is experiencing fatigue. Please generate appropriate feedback for this condition:"

[0363] This will allow all passengers in self-driving vehicles to enjoy travel in a comfortable and safe environment.

[0364] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0365] Step 1:

[0366] When a user speaks or makes an action inside an autonomous vehicle, the infotainment system's camera and microphone capture the user's voice data, facial expression data, and body movement data. Specifically, the microphone picks up the voice signal, and the camera records the facial expression and body movement data in video format. This data is input to the terminal in real time.

[0367] Step 2:

[0368] The device transmits the captured voice data, facial expression data, and body movement data to the server. This transmission is performed simultaneously with the digital conversion of the data. For example, the voice data is sent to the server in WAV format, the facial expression data is sent to the server in image frames, and the body movement data is sent to the server in motion capture information. At this time, the device encodes the data into the appropriate format.

[0369] Step 3:

[0370] The server inputs the received multidimensional data into the emotion engine, which performs facial recognition and expression analysis using OpenCV, audio analysis using PyAudio, and integrates other sensor data. This engine performs tasks such as generating emotion labels (such as "anger," "sadness," and "joy") from facial expression data, and extracting emotional tone, speed, and volume features from audio data.

[0371] Step 4:

[0372] The server passes the emotion information obtained from the emotion engine to the emotion detection model. The emotion detection model uses machine learning frameworks such as TensorFlow and Keras to determine each emotion. It also references past emotion history data, which reconfirms the user's emotional state and improves prediction accuracy. Finally, a highly accurate emotion label is generated, which the server passes to the next step.

[0373] Step 5:

[0374] The server passes the emotion labels obtained from the emotion detection model to the mentoring generative model. A generative AI model (e.g., GPT-2) is used to generate appropriate feedback based on the emotion labels. Specifically, a message such as "Please take an early rest today" is generated from the emotion label "fatigue." The following prompt sentence is used for this generation: "The passenger is feeling fatigued. Please generate appropriate feedback for this state:" The server then passes the generated text data to the next step.

[0375] Step 6:

[0376] The server inputs the generated feedback into a speech synthesis engine, which uses text-to-speech technology to convert the text data into voice data. The converted voice data is then transmitted to the user through the infotainment system. For example, a voice message such as "Please go to bed early today" is played.

[0377] Step 7:

[0378] The device then provides the generated voice data to the user, specifically, voice feedback played through the infotainment system speaker, allowing the user to receive appropriate advice and reassurance.

[0379] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0380] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0381] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0382] [Second embodiment]

[0383] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0384] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0385] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0386] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0387] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0388] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0389] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0390] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0391] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0392] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0393] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0394] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0395] As an embodiment of the present invention, a system using an emotion detection model and a mentoring generation model will be described. This system analyzes a user's voice data in real time and provides appropriate feedback.

[0396] Overall system overview

[0397] The system consists of three main components:

[0398] 1. Emotion detection model

[0399] 2. Mentoring Generative Model

[0400] 3. Speech synthesis engine

[0401] Emotion detection model

[0402] The emotion detection model identifies emotions from a user's voice data. When a user speaks to a device (such as a smart speaker), the voice data is captured and sent to a server. The server inputs this voice data into the emotion detection model and generates an emotion label. The emotion detection model analyzes features such as tone, speed, and volume of the voice to identify the emotion the user is feeling (e.g., anger, sadness, joy, etc.).

[0403] Mentoring Generative Model

[0404] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. The mentoring generative model generates useful feedback for the user according to the identified emotion. For example, if the user is feeling stressed, it generates advice such as "Take a deep breath and relax."

[0405] Text-to-speech engine

[0406] The generated feedback is returned in text format to the server, which then inputs this text into a speech synthesis engine to convert it into a speech format that is easy for the user to understand. The converted speech is then transmitted to the user via the terminal.

[0407] Specific use cases

[0408] 1. Emotion detection and feedback

[0409] 1. The user speaks into the device.

[0410] For example: "I'm really tired today."

[0411] 2. The device captures this audio and sends it to the server as audio data.

[0412] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[0413] 4. The server inputs the generated emotion labels into the mentoring generative model.

[0414] 5. The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as the message "Please take an early rest today."

[0415] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[0416] 7. The device plays the converted audio to the user.

[0417] For example: "Please go to bed early today."

[0418] 8. Users can take appropriate action based on the feedback.

[0419] In this way, users can use the system to quickly detect their emotions and receive appropriate advice and feedback at that moment, which is expected to serve as a useful tool for maintaining mental health.

[0420] The processing flow will be explained below.

[0421] Step 1:

[0422] The user speaks to the device, for example, saying, "I'm really tired today."

[0423] Step 2:

[0424] The device captures this audio and converts it into digital audio data.

[0425] Step 3:

[0426] The terminal transmits digital audio data to the server.

[0427] Step 4:

[0428] The server inputs the received audio data into the emotion detection model.

[0429] Step 5:

[0430] An emotion detection model analyzes the tone, rate, and volume of the speech and generates an emotion label (e.g., "fatigue").

[0431] Step 6:

[0432] The server inputs the generated emotion labels into the mentoring generative model.

[0433] Step 7:

[0434] Based on the emotion labels, the mentoring generative model generates appropriate feedback such as "Please take a rest early today."

[0435] Step 8:

[0436] The server inputs the generated message into a speech synthesis engine and converts it into speech format.

[0437] Step 9:

[0438] The terminal plays back the converted audio data and provides feedback to the user.

[0439] Step 10:

[0440] The user listens to the audio feedback from the device and takes appropriate action, such as taking an early rest.

[0441] Through this series of steps, users' emotions are detected in real time and they can receive appropriate advice on the spot.

[0442] Example 1

[0443] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0444] Conventional emotion detection systems have the drawback of being unable to provide real-time feedback and promptly provide appropriate advice based on the user's emotional state. Another problem is that multiple methods for improving the accuracy of emotion detection and the quality of feedback are not integrated.

[0445] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0446] In this invention, the server includes means for acquiring and transmitting user voice data to the server, means for the server to receive the voice data, input the voice data to an emotion detection model, and analyze the voice features to generate an emotion label, means for the server to input the emotion label to a mentoring generation model and generate an appropriate feedback message, means for inputting the generated feedback message to a voice synthesis engine and converting it into a voice format, and means for transmitting the converted voice to the terminal and playing it back to the user, thereby enabling real-time emotion detection and immediate feedback provision.

[0447] A "user" is a human entity that provides speech data to the system and receives feedback.

[0448] A "terminal" is a device for capturing and transmitting audio data to a server, and is a device for playing back the feedback audio.

[0449] A "server" is a computing system that receives audio data, processes the data using the emotion detection model and the mentoring generation model, and generates feedback.

[0450] An "emotion detection model" is an artificial intelligence model that analyzes characteristics of voice data, such as tone, speed, and volume, to identify a user's emotions.

[0451] An "emotion label" is data that indicates the emotional state of a user extracted from voice data by an emotion detection model.

[0452] The "Mentoring Generative Model" is an artificial intelligence model that generates appropriate feedback messages based on emotion labels.

[0453] A "feedback message" is a text message containing advice or instructions that correspond to the user's emotional state.

[0454] "Speech synthesis engine" means software or a device that converts textual feedback messages into speech.

[0455] "Converted speech" is speech data generated from a text message by a speech synthesis engine.

[0456] "Real-time" refers to processing or feedback that occurs immediately or with very little delay.

[0457] This invention is a system that uses an emotion detection model, a mentoring generation model, and a speech synthesis engine to analyze a user's voice data in real time and provide appropriate feedback. A specific embodiment of this system is described below.

[0458] Overall system configuration

[0459] The system consists of the following main components:

[0460] 1. Emotion detection model

[0461] 2. Mentoring Generative Model

[0462] 3. Speech synthesis engine

[0463] Emotion detection model

[0464] The user provides voice data to the device. For example, the user might say, "I'm really tired today." The device captures the voice data and sends it to the server. The server inputs the received voice data into an emotion detection model. The emotion detection model analyzes features such as tone, speed, and volume of the voice and generates an emotion label. The specific software used is, for example, LibROSA, which extracts the features of the voice data, and a generative AI model based on TensorFlow or PyTorch.

[0465] Mentoring Generative Model

[0466] The server obtains the emotion labels obtained from the emotion detection model and inputs them into the mentoring generative model. The mentoring generative model generates useful feedback for the user based on the identified emotion. For example, if the user is feeling stressed, the mentoring generative model generates a message such as "Please take an early rest today." Generative AI models such as GPT-3 are used here.

[0467] Text-to-speech engine

[0468] The generated feedback message is returned in text format to the server, which then inputs this text into a speech synthesis engine to convert it into a speech format that is easy for the user to understand. Specific software used includes the Google Text-to-Speech API and Amazon Polly. The converted speech is then sent to the device and played back to the user.

[0469] Specific use cases

[0470] A specific example of the system's operation is shown below.

[0471] 1. The user speaks to the device, saying, "I'm really tired today."

[0472] 2. The device captures this audio and sends it to the server as audio data.

[0473] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[0474] 4. The server inputs the generated emotion labels into the mentoring generative model.

[0475] 5. The mentoring generative model generates an appropriate feedback message based on the emotion label "fatigue," such as "Please take an early rest today."

[0476] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[0477] 7. The device plays the converted audio to the user, telling them, "Please go to bed early today."

[0478] 8. Users can take appropriate action based on the feedback.

[0479] A concrete example of a prompt for a generative AI model

[0480] 1. Example prompt for emotion detection model:

[0481] "Based on this audio data, identify the user's current emotion."

[0482] 2. Example prompts for the Mentoring Generative Model:

[0483] "When the emotion label is 'fatigue', generate an appropriate feedback message."

[0484] 3. Example prompt for the speech synthesis engine:

[0485] "Please convert the text 'Please go to bed early today' into audio format."

[0486] In this way, users receive rapid, real-time feedback from the system and can take action to improve their physical and mental health.

[0487] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0488] Step 1:

[0489] The user provides voice data to the terminal, which then captures the voice with a microphone and converts it into digital data.

[0490] Input: User's voice

[0491] Output: Digitized audio data

[0492] Specific operations: Records audio with a microphone, converts the audio data into digital format, and stores the digital data in temporary memory.

[0493] Step 2:

[0494] The device prepares the captured audio data to be sent to the server, dividing it into packets and sending them via TCP / IP.

[0495] Input: Digitized voice data

[0496] Output: Audio data packets to the server

[0497] Specific operation: Packetizes digital data into a specific format and sends it over a network to a server.

[0498] Step 3:

[0499] The server analyzes the received audio data packets and reconstructs the audio data, which is then input into the emotion detection model.

[0500] Input: Voice data packets

[0501] Output: Reconstructed speech data, input data for emotion detection model

[0502] Specific operation: Receive packets, reassemble them appropriately, and combine them into a single audio file. Input the audio into the emotion detection model.

[0503] Step 4:

[0504] The server inputs the audio data into an emotion detection model, analyzes audio features (e.g., tone, rate, volume) and generates emotion labels.

[0505] Input: Reconstructed audio data

[0506] Output: Emotion label

[0507] Specific operation: Extract features from audio data using a library such as LibROSA. Input the extracted features into a generative AI model (e.g., TensorFlow or PyTorch) to obtain emotion labels.

[0508] Step 5:

[0509] The server inputs the emotion labels into a mentoring generative model to generate appropriate feedback messages.

[0510] Input: Emotion label

[0511] Output: Feedback message

[0512] Specific action: The emotion label is input as a prompt sentence into the mentoring generative model to obtain a feedback message (e.g., "Please go to bed early today").

[0513] Step 6:

[0514] The server inputs the generated feedback message into a speech synthesis engine and converts it into speech format.

[0515] Input: Feedback message

[0516] Output: Audio data

[0517] Specific behavior: Converts text messages into speech using the Google Text-to-Speech API, Amazon Polly, etc.

[0518] Step 7:

[0519] The server transmits the feedback message converted into audio format to the terminal.

[0520] Input: Audio data

[0521] Output: Audio data packets

[0522] Specific operation: The generated voice data is packetized for transmission to the terminal and sent via TCP / IP.

[0523] Step 8:

[0524] The device analyzes and reconstructs the received audio data and plays it back through the speaker, thereby providing feedback to the user.

[0525] Input: Voice data packets

[0526] Output: Reconstructed speech data, speech feedback to the user

[0527] Specific operation: The received packets are reconstructed and compiled into a single audio file, which is then played back through the speaker to the user.

[0528] (Application example 1)

[0529] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0530] Conventional food delivery systems have difficulty providing appropriate support and meal suggestions based on the user's emotional state, and have been unable to increase the user's psychological satisfaction. Furthermore, when a user feels fatigued or stressed, there is a lack of a way to provide real-time advice tailored to that emotional state. Therefore, there is a need for improved user experience and psychological health support.

[0531] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0532] In this invention, the server includes means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for communicating the generated feedback to the user by voice, means for including meal suggestions as part of the feedback, and means for including an interface for the user to order food. This makes it possible to provide appropriate feedback and advice according to the user's emotional state, and further to suggest meals and order food delivery according to the emotions.

[0533] An "emotion detection model" is an artificial intelligence model that analyzes a user's voice data and identifies emotions from characteristics such as tone, speed, and volume of the voice.

[0534] The "mentoring generation model" is an artificial intelligence model that generates useful feedback and advice for users based on detected emotions.

[0535] "Speech synthesis engine" refers to software that converts the generated textual feedback into speech.

[0536] "Feedback" refers to appropriate advice or guidance provided based on the user's emotional state.

[0537] "Meal suggestion" refers to selecting and suggesting meals suitable for the user based on the user's emotional state.

[0538] An "interface" refers to the operating screen and input means that a user uses to interact with the system and order food.

[0539] "Voice data" refers to the spoken words or sounds of a user and is digital data captured by a voice recognition system.

[0540] An "emotion label" is an indicator or tag that indicates a user's emotion identified by an emotion detection model.

[0541] As an embodiment of the present invention, a system for analyzing user voice data, detecting emotions, and providing appropriate feedback based on the emotions in a food delivery system will be described. This system mainly consists of the following components:

[0542] 1. Emotion detection model

[0543] 2. Mentoring Generative Model

[0544] 3. Speech synthesis engine

[0545] 4. Meal suggestion tools

[0546] 5. Food ordering interface

[0547] Emotion detection model

[0548] The server receives voice data from users speaking into devices such as smartphones. This voice data is input into an emotion detection model, which analyzes features such as tone, speed, and volume of the voice and generates emotion labels, such as "fatigue" or "joy."

[0549] Mentoring Generative Model

[0550] The server generates appropriate feedback using a mentoring generation model based on the emotion labels obtained from the emotion detection model. This feedback includes advice and guidance based on the emotion the user is feeling. For example, if the user has an emotion label of "fatigue," feedback including a specific meal suggestion such as "hot soup would be good" is generated.

[0551] Text-to-speech engine

[0552] The generated feedback is sent from the server to a speech synthesis engine, which converts the feedback text into speech, which is then transmitted to the user via the terminal.

[0553] Meal suggestion tool

[0554] The feedback generated by the mentoring generative model includes specific meal suggestions based on the user's emotional state, which are structured to help the user choose appropriate meals.

[0555] Food Ordering Interface

[0556] An interface is provided for the user to order food based on the suggested meal menu, and the user can easily use this interface to order the suggested meal.

[0557] Specific examples

[0558] For example, suppose a user says to the device, "I'm tired today, so I want to eat something comforting." The device captures the voice data and sends it to the emotion detection model. The server obtains the emotion label "fatigue" and uses the mentoring generative model to generate feedback such as "If you're feeling tired, a warm soup would be good." This feedback is sent to the speech synthesis engine, converted into audio format, and transmitted to the user. The user can easily order the suggested meal through the provided food ordering interface.

[0559] Prompt Sentence Examples

[0560] "User voice text: 'I'm tired today and I want to eat something comforting.'

[0561] Emotion detection model output: 'Fatigue'

[0562] Generate mentoring messages based on this emotion.”

[0563] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0564] Step 1:

[0565] The user speaks to the device and inputs voice data. For example, if the user says, "I'm tired today, so I want to eat something soothing," the voice is captured by the device.

[0566] Step 2:

[0567] The device sends the captured audio data to the server, which then becomes the input data for the system.

[0568] Step 3:

[0569] The server inputs the received voice data into an emotion detection model, which analyzes the tone, speed, volume, etc. of the voice to identify the user's emotion label. In this step, an emotion label such as "fatigue" is output.

[0570] Step 4:

[0571] The server inputs the generated emotion labels into the mentoring generative model, which then generates an appropriate feedback message based on the emotion labels. For example, for the emotion label "fatigue," the model generates the message "If you're feeling tired, warm soup would be good."

[0572] Step 5:

[0573] The server inputs the generated feedback message into a speech synthesis engine, which converts the text message into a voice message. Specifically, the text "If you're feeling tired, a warm soup would be good" is converted into voice data.

[0574] Step 6:

[0575] The server transmits the converted voice data to the terminal, which then plays the voice data and transmits it to the user, so that the user receives voice feedback.

[0576] Step 7:

[0577] The server generates a message containing specific meal suggestions as part of the feedback, such as a specific menu item like "hot soup," in a format that makes it easy for the user to choose.

[0578] Step 8:

[0579] The terminal provides a user interface that displays the suggested meal menu and provides a means for the user to perform operations to order it. The user can use this interface to easily order the suggested meal.

[0580] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0581] As an embodiment of the present invention, we will explain a system that combines an emotion engine that recognizes a user's emotions using information such as facial expressions and body movements in addition to voice data. This system quickly detects a user's diverse emotional information and provides appropriate feedback in real time.

[0582] Overall system overview

[0583] The system consists of four main components:

[0584] 1. Emotion Engine

[0585] 2. Emotion Detection Model

[0586] 3. Mentoring Generative Model

[0587] 4. Speech synthesis engine

[0588] Emotion Engine

[0589] The emotion engine recognizes emotions from the user's voice data, facial expressions, and body movements. When the user is close to the device, the device's camera and sensors capture facial expressions and body movements. In addition, voice data is also acquired. This data is sent to the server and input into the emotion engine. The emotion engine integrates this multidimensional data and recognizes the user's emotions (e.g., joy, anger, sadness, surprise, etc.) with high accuracy.

[0590] Emotion detection model

[0591] The emotion detection model further analyzes the voice data based on the emotional information provided by the emotion engine. This model further analyzes voice characteristics such as tone, speed, and volume to determine the emotion. It can also refer to the user's past emotional history to predict the user's current emotional state.

[0592] Mentoring Generative Model

[0593] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. This model generates advice according to the identified emotion and provides useful feedback to the user.

[0594] Text-to-speech engine

[0595] The generated feedback is returned to the server in text format and converted into audio format by inputting it into a speech synthesis engine. The device plays back this audio data and conveys the feedback to the user.

[0596] Specific use cases

[0597] 1. Recognizing emotions and providing feedback

[0598] 1. As the user speaks to the device, data including facial expressions and body movements is captured by the device's camera and sensors.

[0599] For example: "I'm really tired today."

[0600] 2. The device transmits this voice data, facial expression data, and body movement data to the server.

[0601] 3. The server inputs the received multidimensional data into the emotion engine, which analyzes the data and generates an emotion label, such as "fatigue."

[0602] 4. The server re-inputs the audio data, along with the emotion labels obtained from the emotion engine, into the emotion detection model, which analyzes the audio features and reconfirms the results from the emotion engine.

[0603] 5. The server requests the mentoring model to generate feedback based on the determined emotion label. In response to the emotion "fatigue," the mentoring model generates advice such as "Take an early rest today."

[0604] 6. The server inputs this advice into a speech synthesis engine and converts it into voice data.

[0605] 7. The device plays the generated audio data and provides feedback to the user.

[0606] For example: "Please go to bed early today."

[0607] 8. The user follows the feedback and takes appropriate action.

[0608] This system allows users to receive accurate feedback in real time based on a variety of emotional data, which is expected to provide more effective support for users' mental health.

[0609] The processing flow will be explained below.

[0610] Step 1:

[0611] As a user speaks to the device, the device's camera and sensors capture the user's facial expressions and body movements. For example, the user might say, "I'm really tired today."

[0612] Step 2:

[0613] The device digitizes the captured voice data, facial expression data, and body movement data and transmits them to a server.

[0614] Step 3:

[0615] The server inputs the received voice data, facial expression data, and body movement data into the emotion engine, which integrates and analyzes the data to identify the user's emotional state.

[0616] Step 4:

[0617] Based on the analysis results, the emotion engine generates an emotion label called "fatigue" and returns it to the server.

[0618] Step 5:

[0619] The server inputs the emotion labels and audio data obtained from the emotion engine into the emotion detection model, which analyzes the tone, rate, and volume of the audio and double-checks the results from the emotion engine.

[0620] Step 6:

[0621] The emotion detection model determines the emotion "fatigue" and returns the emotion label to the server.

[0622] Step 7:

[0623] The server inputs the determined emotion labels into the mentoring generative model.

[0624] Step 8:

[0625] The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as "Please take an early rest today."

[0626] Step 9:

[0627] The server inputs the generated feedback into a speech synthesis engine, which converts the text-format feedback into speech data.

[0628] Step 10:

[0629] The terminal plays back the converted audio data and provides feedback to the user.

[0630] Step 11:

[0631] The user listens to the audio feedback from the device and takes appropriate action based on it, such as taking an early rest.

[0632] Through this series of steps, users can have their emotions detected from multiple angles using the emotion engine, and receive optimal feedback in real time.

[0633] Example 2

[0634] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0635] Conventional emotion recognition systems detect emotions based solely on the user's voice data, resulting in insufficient emotion recognition accuracy. Furthermore, the quality and applicability of the feedback provided is limited, resulting in insufficient user response. Therefore, there is a need for a system that can improve emotion recognition accuracy and provide more appropriate feedback.

[0636] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0637] In this invention, the server includes means for recognizing emotions from multiple data sources including facial expressions and body movements, means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, and means for transmitting the generated feedback to the user by voice, thereby improving the accuracy of emotion recognition and enabling more appropriate and useful feedback to be provided to the user in real time.

[0638] - "Multiple data sources including facial expressions and body movements" is a general term for input methods such as sensors, cameras, and microphones that collect multiple types of data, such as a user's facial expressions, body movements, and even voice data.

[0639] "Means for recognizing emotions" refers to a system or algorithm that analyzes collected voice data, facial expression data, and body movement data to identify a user's emotions.

[0640] An "emotion detection model" refers to a machine learning model or algorithm that analyzes a user's voice data and detects emotions from its characteristics, such as tone, speed, and volume.

[0641] "Mentoring generative model" means a system or algorithm that generates appropriate feedback and advice for a user based on emotional information obtained from an emotion detection model.

[0642] "Means for generating feedback" refers to software or a program that uses an emotion detection model or a mentoring generation model to generate advice or guidance according to the user's emotions.

[0643] "Means for transmitting the generated feedback to the user by voice" refers to a system that uses a speech synthesis engine to convert text-based feedback into audio format and transmits it to the user through speakers or earphones.

[0644] An embodiment of the present invention is a system that recognizes emotions using a user's voice data, facial expressions, and body movements, and provides appropriate feedback in real time.

[0645] System Overview

[0646] The system mainly consists of the following elements:

[0647] 1. Data sources including facial expressions and body movements

[0648] 2. Means of recognizing emotions (emotion engine)

[0649] 3. Emotion Detection Model

[0650] 4. Mentoring Generative Model

[0651] 5. Speech synthesis engine

[0652] System configuration

[0653] 1. Data Source:

[0654] The user speaks to the device, and the device's built-in camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data.

[0655] As a specific example, consider the case where a user says, "I'm really tired today." At this time, the device captures the user's tired facial expression, slumped shoulders, and other aspects of their posture, and the microphone records their voice.

[0656] 2. Means of Emotion Recognition (Emotion Engine):

[0657] The device sends the collected data to a server in real time. The emotion engine on the server analyzes this multidimensional data and recognizes the user's emotions. Specifically, it uses facial expression recognition algorithms and motion analysis algorithms.

[0658] 3. Emotion detection model:

[0659] The server inputs the voice data into an emotion detection model, which analyzes voice characteristics such as tone, speed, and volume to detect emotions. It also uses past emotion data to improve the accuracy of emotion detection.

[0660] 4. Mentoring generative model:

[0661] The server receives the output of the emotion detection model and generates appropriate feedback using the mentoring generation model. The generated feedback includes advice and comments according to the user's emotional state. For example, the generated feedback might be "Please go to bed early today."

[0662] 5. Speech synthesis engine:

[0663] The server inputs the generated feedback into a speech synthesis engine, converting the text into natural-sounding speech data, which is then sent to the device.

[0664] 6. Providing feedback:

[0665] The device plays back the voice data sent from the speech synthesis engine and provides feedback to the user. For example, a voice message saying "Please go to bed early today" is played through the speaker or earphones.

[0666] Prompt Sentence Examples

[0667] Below are some example prompts to ensure the system works properly:

[0668] Capture the voice data, facial expressions, and body movements of a user saying, "I'm really tired today," and input them into the emotion engine. After that, once the emotion label "fatigue" is generated, reanalyze it with the emotion detection model, and finally generate the advice "Please take an early rest today," and convert it into voice data with the speech synthesis engine.

[0669] This system recognizes the user's emotions with high accuracy from multiple data sources, including not only their voice but also their facial expressions and body movements, and provides real-time feedback according to their current state, making it possible to more appropriately support the user's physical and mental state.

[0670] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0671] Step 1:

[0672] The user speaks to the device, and the device's camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data (input: user's voice, facial expressions, and body movements; output: multiple data sources).

[0673] Step 2:

[0674] The device transmits the captured facial expression data, body movement data, and voice data to the server in real time (input: multiple data sources, output: data transfer to the server).

[0675] Step 3:

[0676] The server inputs the received multidimensional data into the emotion engine, which then integrates the data and identifies the user's emotion using facial expression recognition and motion analysis algorithms (input: data transferred to the server, output: recognized emotion label).

[0677] Step 4:

[0678] The server inputs the emotion label and audio data output from the emotion engine into the emotion detection model, which analyzes features such as tone, speed, and volume of the audio to re-detect the emotion in detail (input: emotion label, audio data, output: confirmed emotion label).

[0679] Step 5:

[0680] The server inputs the determined emotion label into the mentoring generative model and generates feedback. The mentoring generative model generates appropriate advice based on the identified emotion (input: determined emotion label, output: generated feedback).

[0681] Step 6:

[0682] The server inputs the generated feedback into a speech synthesis engine and converts the text feedback into speech data (input: generated feedback text, output: speech data).

[0683] Step 7:

[0684] The device receives the voice data sent from the server and provides feedback to the user through a speaker or earphones (input: voice data, output: voice message to the user).

[0685] Specific operation example

[0686] 1. The user says to the device, "I'm really tired today."

[0687] 2. The device collects the user's facial expressions, body movements, and voice data and sends it to the server.

[0688] 3. The server uses an emotion engine to recognize the emotion "fatigue."

[0689] 4. The server performs detailed analysis of the voice data using an emotion detection model to reconfirm the emotion.

[0690] 5. The server uses a mentoring generation model to generate advice such as "Take a rest early today," and converts it into voice data using a speech synthesis engine.

[0691] 6. The device plays the generated voice data and provides the user with feedback such as "Please go to bed early today."

[0692] (Application example 2)

[0693] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0694] In autonomous vehicles, the challenge is to monitor the emotional state of passengers in real time and improve their comfort and safety. This will provide an environment where passengers can use autonomous vehicles safely and without anxiety or stress, even for long periods of time.

[0695] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0696] In this invention, the server includes means for detecting emotions from the user's voice data, facial expression data, and body movement data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for providing the feedback to the infotainment system of the autonomous vehicle, and means for communicating the generated feedback to the user by voice. This makes it possible to recognize the passenger's emotional state in real time and provide advice or adjust environmental settings (music, lighting, interior temperature) according to the state.

[0697] An "emotion detection model" is a system for detecting emotions by analyzing a user's voice data, facial expression data, and body movement data.

[0698] The "mentoring generation model" is a system that generates appropriate feedback and advice based on emotions detected by the emotion detection model.

[0699] An "infotainment system" is a system that provides information and entertainment functions and is installed in an autonomous vehicle.

[0700] "Feedback" refers to advice and information generated by the emotion detection model and mentoring generation model according to the user's emotional state.

[0701] "Voice data" is a digital representation of a voice signal containing the content of a user's speech.

[0702] "Facial expression data" refers to a facial expression of a user captured by a photographing device such as a camera and expressed in digital form.

[0703] "Body movement data" refers to the user's body movements captured by sensors or other devices and expressed in digital form.

[0704] An "automated vehicle" is a vehicle whose driving operation is performed by an automatic control system.

[0705] This paper describes an embodiment of the present invention. The system analyzes a user's voice data, facial expression data, and body movements to recognize emotions and improve passenger comfort and safety in an autonomous vehicle. Specifically, it consists of the following four main components:

[0706] 1. Emotion Engine

[0707] 2. Emotion Detection Model

[0708] 3. Mentoring Generative Model

[0709] 4. Speech synthesis engine

[0710] Emotion Engine

[0711] The emotion engine captures the user's voice data, facial expression data, and body movement data. When a user gets into an autonomous vehicle, cameras and sensors installed in the vehicle capture this data and send it to a server. This uses tools such as PyAudio to capture voice data and OpenCV to capture facial expressions and body movements. The data sent to the server is analyzed as multidimensional data by the emotion engine to recognize the user's emotions.

[0712] Emotion detection model

[0713] The emotion detection model uses the emotional information provided by the emotion engine to perform detailed analysis of voice characteristics such as tone, speed, volume, facial expressions, and body movements. This allows it to accurately determine the user's emotional state. The model is built using machine learning frameworks such as TensorFlow and Keras. For example, if the user is relaxed, it generates a label such as "calm," and if they are nervous, it generates a label such as "nervous."

[0714] Mentoring Generative Model

[0715] The mentoring generative model generates appropriate feedback based on the emotion labels obtained from the emotion detection model. The server inputs the emotion labels received from the emotion detection model into a generative AI model (e.g., GPT-2) and generates feedback in text format. This feedback may include advice based on the passenger's emotions or adjustments to environmental settings (music, lighting, interior temperature, etc.).

[0716] Text-to-speech engine

[0717] The generated feedback is input into a speech synthesis engine and converted into voice format. The server then transmits this voice data to the passenger through the infotainment system in the autonomous vehicle. Specific examples of feedback could be, "Take a short break today" or "We'll lower the temperature inside the car a little."

[0718] Specific use cases

[0719] For example, if a user in an autonomous vehicle says, "Oh, I'm really tired today," a camera and microphone installed in the infotainment system capture their voice, facial expressions, and body movements. The voice and facial expression data are sent to a server and analyzed by an emotion engine. If the emotion detection model determines that the user is "fatigue," the mentoring generation model generates feedback such as "Please take an early rest today," which is then converted into voice data by a speech synthesis engine and provided to the user. This feedback contributes to creating an environment where the user can feel safe and relaxed.

[0720] Example prompt sentence:

[0721] "The passenger is experiencing fatigue. Please generate appropriate feedback for this condition:"

[0722] This will allow all passengers in self-driving vehicles to enjoy travel in a comfortable and safe environment.

[0723] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0724] Step 1:

[0725] When a user speaks or makes an action inside an autonomous vehicle, the infotainment system's camera and microphone capture the user's voice data, facial expression data, and body movement data. Specifically, the microphone picks up the voice signal, and the camera records the facial expression and body movement data in video format. This data is input to the terminal in real time.

[0726] Step 2:

[0727] The device transmits the captured voice data, facial expression data, and body movement data to the server. This transmission is performed simultaneously with the digital conversion of the data. For example, the voice data is sent to the server in WAV format, the facial expression data is sent to the server in image frames, and the body movement data is sent to the server in motion capture information. At this time, the device encodes the data into the appropriate format.

[0728] Step 3:

[0729] The server inputs the received multidimensional data into the emotion engine, which performs facial recognition and expression analysis using OpenCV, audio analysis using PyAudio, and integrates other sensor data. This engine performs tasks such as generating emotion labels (such as "anger," "sadness," and "joy") from facial expression data, and extracting emotional tone, speed, and volume features from audio data.

[0730] Step 4:

[0731] The server passes the emotion information obtained from the emotion engine to the emotion detection model. The emotion detection model uses machine learning frameworks such as TensorFlow and Keras to determine each emotion. It also references past emotion history data, which reconfirms the user's emotional state and improves prediction accuracy. Finally, a highly accurate emotion label is generated, which the server passes to the next step.

[0732] Step 5:

[0733] The server passes the emotion labels obtained from the emotion detection model to the mentoring generative model. A generative AI model (e.g., GPT-2) is used to generate appropriate feedback based on the emotion labels. Specifically, a message such as "Please take an early rest today" is generated from the emotion label "fatigue." The following prompt sentence is used for this generation: "The passenger is feeling fatigued. Please generate appropriate feedback for this state:" The server then passes the generated text data to the next step.

[0734] Step 6:

[0735] The server inputs the generated feedback into a speech synthesis engine, which uses text-to-speech technology to convert the text data into voice data. The converted voice data is then transmitted to the user through the infotainment system. For example, a voice message such as "Please go to bed early today" is played.

[0736] Step 7:

[0737] The device then provides the generated voice data to the user, specifically, voice feedback played through the infotainment system speaker, allowing the user to receive appropriate advice and reassurance.

[0738] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0739] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0740] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0741] [Third embodiment]

[0742] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0743] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0744] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0745] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0746] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0747] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0748] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0749] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0750] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0751] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0752] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0753] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0754] As an embodiment of the present invention, a system using an emotion detection model and a mentoring generation model will be described. This system analyzes a user's voice data in real time and provides appropriate feedback.

[0755] Overall system overview

[0756] The system consists of three main components:

[0757] 1. Emotion detection model

[0758] 2. Mentoring Generative Model

[0759] 3. Speech synthesis engine

[0760] Emotion detection model

[0761] The emotion detection model identifies emotions from a user's voice data. When a user speaks to a device (such as a smart speaker), the voice data is captured and sent to a server. The server inputs this voice data into the emotion detection model and generates an emotion label. The emotion detection model analyzes features such as tone, speed, and volume of the voice to identify the emotion the user is feeling (e.g., anger, sadness, joy, etc.).

[0762] Mentoring Generative Model

[0763] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. The mentoring generative model generates useful feedback for the user according to the identified emotion. For example, if the user is feeling stressed, it generates advice such as "Take a deep breath and relax."

[0764] Text-to-speech engine

[0765] The generated feedback is returned in text format to the server, which then inputs this text into a speech synthesis engine to convert it into a speech format that is easy for the user to understand. The converted speech is then transmitted to the user via the terminal.

[0766] Specific use cases

[0767] 1. Emotion detection and feedback

[0768] 1. The user speaks into the device.

[0769] For example: "I'm really tired today."

[0770] 2. The device captures this audio and sends it to the server as audio data.

[0771] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[0772] 4. The server inputs the generated emotion labels into the mentoring generative model.

[0773] 5. The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as the message "Please take an early rest today."

[0774] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[0775] 7. The device plays the converted audio to the user.

[0776] For example: "Please go to bed early today."

[0777] 8. Users can take appropriate action based on the feedback.

[0778] In this way, users can use the system to quickly detect their emotions and receive appropriate advice and feedback at that moment, which is expected to serve as a useful tool for maintaining mental health.

[0779] The processing flow will be explained below.

[0780] Step 1:

[0781] The user speaks to the device, for example, saying, "I'm really tired today."

[0782] Step 2:

[0783] The device captures this audio and converts it into digital audio data.

[0784] Step 3:

[0785] The terminal transmits digital audio data to the server.

[0786] Step 4:

[0787] The server inputs the received audio data into the emotion detection model.

[0788] Step 5:

[0789] An emotion detection model analyzes the tone, rate, and volume of the speech and generates an emotion label (e.g., "fatigue").

[0790] Step 6:

[0791] The server inputs the generated emotion labels into the mentoring generative model.

[0792] Step 7:

[0793] Based on the emotion labels, the mentoring generative model generates appropriate feedback such as "Please take a rest early today."

[0794] Step 8:

[0795] The server inputs the generated message into a speech synthesis engine and converts it into speech format.

[0796] Step 9:

[0797] The terminal plays back the converted audio data and provides feedback to the user.

[0798] Step 10:

[0799] The user listens to the audio feedback from the device and takes appropriate action, such as taking an early rest.

[0800] Through this series of steps, users' emotions are detected in real time and they can receive appropriate advice on the spot.

[0801] Example 1

[0802] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0803] Conventional emotion detection systems have the drawback of being unable to provide real-time feedback and promptly provide appropriate advice based on the user's emotional state. Another problem is that multiple methods for improving the accuracy of emotion detection and the quality of feedback are not integrated.

[0804] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0805] In this invention, the server includes means for acquiring and transmitting user voice data to the server, means for the server to receive the voice data, input the voice data to an emotion detection model, and analyze the voice features to generate an emotion label, means for the server to input the emotion label to a mentoring generation model and generate an appropriate feedback message, means for inputting the generated feedback message to a voice synthesis engine and converting it into a voice format, and means for transmitting the converted voice to the terminal and playing it back to the user, thereby enabling real-time emotion detection and immediate feedback provision.

[0806] A "user" is a human entity that provides speech data to the system and receives feedback.

[0807] A "terminal" is a device for capturing and transmitting audio data to a server, and is a device for playing back the feedback audio.

[0808] A "server" is a computing system that receives audio data, processes the data using the emotion detection model and the mentoring generation model, and generates feedback.

[0809] An "emotion detection model" is an artificial intelligence model that analyzes characteristics of voice data, such as tone, speed, and volume, to identify a user's emotions.

[0810] An "emotion label" is data that indicates the emotional state of a user extracted from voice data by an emotion detection model.

[0811] The "Mentoring Generative Model" is an artificial intelligence model that generates appropriate feedback messages based on emotion labels.

[0812] A "feedback message" is a text message containing advice or instructions that correspond to the user's emotional state.

[0813] "Speech synthesis engine" means software or a device that converts textual feedback messages into speech.

[0814] "Converted speech" is speech data generated from a text message by a speech synthesis engine.

[0815] "Real-time" refers to processing or feedback that occurs immediately or with very little delay.

[0816] This invention is a system that uses an emotion detection model, a mentoring generation model, and a speech synthesis engine to analyze a user's voice data in real time and provide appropriate feedback. A specific embodiment of this system is described below.

[0817] Overall system configuration

[0818] The system consists of the following main components:

[0819] 1. Emotion detection model

[0820] 2. Mentoring Generative Model

[0821] 3. Speech synthesis engine

[0822] Emotion detection model

[0823] The user provides voice data to the device. For example, the user might say, "I'm really tired today." The device captures the voice data and sends it to the server. The server inputs the received voice data into an emotion detection model. The emotion detection model analyzes features such as tone, speed, and volume of the voice and generates an emotion label. The specific software used is, for example, LibROSA, which extracts the features of the voice data, and a generative AI model based on TensorFlow or PyTorch.

[0824] Mentoring Generative Model

[0825] The server obtains the emotion labels obtained from the emotion detection model and inputs them into the mentoring generative model. The mentoring generative model generates useful feedback for the user based on the identified emotion. For example, if the user is feeling stressed, the mentoring generative model generates a message such as "Please take an early rest today." Generative AI models such as GPT-3 are used here.

[0826] Text-to-speech engine

[0827] The generated feedback message is returned in text format to the server, which then inputs this text into a speech synthesis engine to convert it into a speech format that is easy for the user to understand. Specific software used includes the Google Text-to-Speech API and Amazon Polly. The converted speech is then sent to the device and played back to the user.

[0828] Specific use cases

[0829] A specific example of the system's operation is shown below.

[0830] 1. The user speaks to the device, saying, "I'm really tired today."

[0831] 2. The device captures this audio and sends it to the server as audio data.

[0832] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[0833] 4. The server inputs the generated emotion labels into the mentoring generative model.

[0834] 5. The mentoring generative model generates an appropriate feedback message based on the emotion label "fatigue," such as "Please take an early rest today."

[0835] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[0836] 7. The device plays the converted audio to the user, telling them, "Please go to bed early today."

[0837] 8. Users can take appropriate action based on the feedback.

[0838] A concrete example of a prompt for a generative AI model

[0839] 1. Example prompt for emotion detection model:

[0840] "Based on this audio data, identify the user's current emotion."

[0841] 2. Example prompts for the Mentoring Generative Model:

[0842] "When the emotion label is 'fatigue', generate an appropriate feedback message."

[0843] 3. Example prompt for the speech synthesis engine:

[0844] "Please convert the text 'Please go to bed early today' into audio format."

[0845] In this way, users receive rapid, real-time feedback from the system and can take action to improve their physical and mental health.

[0846] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0847] Step 1:

[0848] The user provides voice data to the terminal, which then captures the voice with a microphone and converts it into digital data.

[0849] Input: User's voice

[0850] Output: Digitized audio data

[0851] Specific operations: Records audio with a microphone, converts the audio data into digital format, and stores the digital data in temporary memory.

[0852] Step 2:

[0853] The device prepares the captured audio data to be sent to the server, dividing it into packets and sending them via TCP / IP.

[0854] Input: Digitized voice data

[0855] Output: Audio data packets to the server

[0856] Specific operation: Packetizes digital data into a specific format and sends it over a network to a server.

[0857] Step 3:

[0858] The server analyzes the received audio data packets and reconstructs the audio data, which is then input into the emotion detection model.

[0859] Input: Voice data packets

[0860] Output: Reconstructed speech data, input data for emotion detection model

[0861] Specific operation: Receive packets, reassemble them appropriately, and combine them into a single audio file. Input the audio into the emotion detection model.

[0862] Step 4:

[0863] The server inputs the audio data into an emotion detection model, analyzes audio features (e.g., tone, rate, volume) and generates emotion labels.

[0864] Input: Reconstructed audio data

[0865] Output: Emotion label

[0866] Specific operation: Extract features from audio data using a library such as LibROSA. Input the extracted features into a generative AI model (e.g., TensorFlow or PyTorch) to obtain emotion labels.

[0867] Step 5:

[0868] The server inputs the emotion labels into a mentoring generative model to generate appropriate feedback messages.

[0869] Input: Emotion label

[0870] Output: Feedback message

[0871] Specific action: The emotion label is input as a prompt sentence into the mentoring generative model to obtain a feedback message (e.g., "Please go to bed early today").

[0872] Step 6:

[0873] The server inputs the generated feedback message into a speech synthesis engine and converts it into speech format.

[0874] Input: Feedback message

[0875] Output: Audio data

[0876] Specific behavior: Converts text messages into speech using the Google Text-to-Speech API, Amazon Polly, etc.

[0877] Step 7:

[0878] The server transmits the feedback message converted into audio format to the terminal.

[0879] Input: Audio data

[0880] Output: Audio data packets

[0881] Specific operation: The generated voice data is packetized for transmission to the terminal and sent via TCP / IP.

[0882] Step 8:

[0883] The device analyzes and reconstructs the received audio data and plays it back through the speaker, thereby providing feedback to the user.

[0884] Input: Voice data packets

[0885] Output: Reconstructed speech data, speech feedback to the user

[0886] Specific operation: The received packets are reconstructed and compiled into a single audio file, which is then played back through the speaker to the user.

[0887] (Application example 1)

[0888] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0889] Conventional food delivery systems have difficulty providing appropriate support and meal suggestions based on the user's emotional state, and have been unable to increase the user's psychological satisfaction. Furthermore, when a user feels fatigued or stressed, there is a lack of a way to provide real-time advice tailored to that emotional state. Therefore, there is a need for improved user experience and psychological health support.

[0890] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0891] In this invention, the server includes means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for communicating the generated feedback to the user by voice, means for including meal suggestions as part of the feedback, and means for including an interface for the user to order food. This makes it possible to provide appropriate feedback and advice according to the user's emotional state, and further to suggest meals and order food delivery according to the emotions.

[0892] An "emotion detection model" is an artificial intelligence model that analyzes a user's voice data and identifies emotions from characteristics such as tone, speed, and volume of the voice.

[0893] The "mentoring generation model" is an artificial intelligence model that generates useful feedback and advice for users based on detected emotions.

[0894] "Speech synthesis engine" refers to software that converts the generated textual feedback into speech.

[0895] "Feedback" refers to appropriate advice or guidance provided based on the user's emotional state.

[0896] "Meal suggestion" refers to selecting and suggesting meals suitable for the user based on the user's emotional state.

[0897] An "interface" refers to the operating screen and input means that a user uses to interact with the system and order food.

[0898] "Voice data" refers to the spoken words or sounds of a user and is digital data captured by a voice recognition system.

[0899] An "emotion label" is an indicator or tag that indicates a user's emotion identified by an emotion detection model.

[0900] As an embodiment of the present invention, a system for analyzing user voice data, detecting emotions, and providing appropriate feedback based on the emotions in a food delivery system will be described. This system mainly consists of the following components:

[0901] 1. Emotion detection model

[0902] 2. Mentoring Generative Model

[0903] 3. Speech synthesis engine

[0904] 4. Meal suggestion tools

[0905] 5. Food ordering interface

[0906] Emotion detection model

[0907] The server receives voice data from users speaking into devices such as smartphones. This voice data is input into an emotion detection model, which analyzes features such as tone, speed, and volume of the voice and generates emotion labels, such as "fatigue" or "joy."

[0908] Mentoring Generative Model

[0909] The server generates appropriate feedback using a mentoring generation model based on the emotion labels obtained from the emotion detection model. This feedback includes advice and guidance based on the emotion the user is feeling. For example, if the user has an emotion label of "fatigue," feedback including a specific meal suggestion such as "hot soup would be good" is generated.

[0910] Text-to-speech engine

[0911] The generated feedback is sent from the server to a speech synthesis engine, which converts the feedback text into speech, which is then transmitted to the user via the terminal.

[0912] Meal suggestion tool

[0913] The feedback generated by the mentoring generative model includes specific meal suggestions based on the user's emotional state, which are structured to help the user choose appropriate meals.

[0914] Food Ordering Interface

[0915] An interface is provided for the user to order food based on the suggested meal menu, and the user can easily use this interface to order the suggested meal.

[0916] Specific examples

[0917] For example, suppose a user says to the device, "I'm tired today, so I want to eat something comforting." The device captures the voice data and sends it to the emotion detection model. The server obtains the emotion label "fatigue" and uses the mentoring generative model to generate feedback such as "If you're feeling tired, a warm soup would be good." This feedback is sent to the speech synthesis engine, converted into audio format, and transmitted to the user. The user can easily order the suggested meal through the provided food ordering interface.

[0918] Prompt Sentence Examples

[0919] "User voice text: 'I'm tired today and I want to eat something comforting.'

[0920] Emotion detection model output: 'Fatigue'

[0921] Generate mentoring messages based on this emotion.”

[0922] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0923] Step 1:

[0924] The user speaks to the device and inputs voice data. For example, if the user says, "I'm tired today, so I want to eat something soothing," the voice is captured by the device.

[0925] Step 2:

[0926] The device sends the captured audio data to the server, which then becomes the input data for the system.

[0927] Step 3:

[0928] The server inputs the received voice data into an emotion detection model, which analyzes the tone, speed, volume, etc. of the voice to identify the user's emotion label. In this step, an emotion label such as "fatigue" is output.

[0929] Step 4:

[0930] The server inputs the generated emotion labels into the mentoring generative model, which then generates an appropriate feedback message based on the emotion labels. For example, for the emotion label "fatigue," the model generates the message "If you're feeling tired, warm soup would be good."

[0931] Step 5:

[0932] The server inputs the generated feedback message into a speech synthesis engine, which converts the text message into a voice message. Specifically, the text "If you're feeling tired, a warm soup would be good" is converted into voice data.

[0933] Step 6:

[0934] The server transmits the converted voice data to the terminal, which then plays the voice data and transmits it to the user, so that the user receives voice feedback.

[0935] Step 7:

[0936] The server generates a message containing specific meal suggestions as part of the feedback, such as a specific menu item like "hot soup," in a format that makes it easy for the user to choose.

[0937] Step 8:

[0938] The terminal provides a user interface that displays the suggested meal menu and provides a means for the user to perform operations to order it. The user can use this interface to easily order the suggested meal.

[0939] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0940] As an embodiment of the present invention, we will explain a system that combines an emotion engine that recognizes a user's emotions using information such as facial expressions and body movements in addition to voice data. This system quickly detects a user's diverse emotional information and provides appropriate feedback in real time.

[0941] Overall system overview

[0942] The system consists of four main components:

[0943] 1. Emotion Engine

[0944] 2. Emotion Detection Model

[0945] 3. Mentoring Generative Model

[0946] 4. Speech synthesis engine

[0947] Emotion Engine

[0948] The emotion engine recognizes emotions from the user's voice data, facial expressions, and body movements. When the user is close to the device, the device's camera and sensors capture facial expressions and body movements. In addition, voice data is also acquired. This data is sent to the server and input into the emotion engine. The emotion engine integrates this multidimensional data and recognizes the user's emotions (e.g., joy, anger, sadness, surprise, etc.) with high accuracy.

[0949] Emotion detection model

[0950] The emotion detection model further analyzes the voice data based on the emotional information provided by the emotion engine. This model further analyzes voice characteristics such as tone, speed, and volume to determine the emotion. It can also refer to the user's past emotional history to predict the user's current emotional state.

[0951] Mentoring Generative Model

[0952] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. This model generates advice according to the identified emotion and provides useful feedback to the user.

[0953] Text-to-speech engine

[0954] The generated feedback is returned to the server in text format and converted into audio format by inputting it into a speech synthesis engine. The device plays back this audio data and conveys the feedback to the user.

[0955] Specific use cases

[0956] 1. Recognizing emotions and providing feedback

[0957] 1. As the user speaks to the device, data including facial expressions and body movements is captured by the device's camera and sensors.

[0958] For example: "I'm really tired today."

[0959] 2. The device transmits this voice data, facial expression data, and body movement data to the server.

[0960] 3. The server inputs the received multidimensional data into the emotion engine, which analyzes the data and generates an emotion label, such as "fatigue."

[0961] 4. The server re-inputs the audio data, along with the emotion labels obtained from the emotion engine, into the emotion detection model, which analyzes the audio features and reconfirms the results from the emotion engine.

[0962] 5. The server requests the mentoring model to generate feedback based on the determined emotion label. In response to the emotion "fatigue," the mentoring model generates advice such as "Take an early rest today."

[0963] 6. The server inputs this advice into a speech synthesis engine and converts it into voice data.

[0964] 7. The device plays the generated audio data and provides feedback to the user.

[0965] For example: "Please go to bed early today."

[0966] 8. The user follows the feedback and takes appropriate action.

[0967] This system allows users to receive accurate feedback in real time based on a variety of emotional data, which is expected to provide more effective support for users' mental health.

[0968] The processing flow will be explained below.

[0969] Step 1:

[0970] As a user speaks to the device, the device's camera and sensors capture the user's facial expressions and body movements. For example, the user might say, "I'm really tired today."

[0971] Step 2:

[0972] The device digitizes the captured voice data, facial expression data, and body movement data and transmits them to a server.

[0973] Step 3:

[0974] The server inputs the received voice data, facial expression data, and body movement data into the emotion engine, which integrates and analyzes the data to identify the user's emotional state.

[0975] Step 4:

[0976] Based on the analysis results, the emotion engine generates an emotion label called "fatigue" and returns it to the server.

[0977] Step 5:

[0978] The server inputs the emotion labels and audio data obtained from the emotion engine into the emotion detection model, which analyzes the tone, rate, and volume of the audio and double-checks the results from the emotion engine.

[0979] Step 6:

[0980] The emotion detection model determines the emotion "fatigue" and returns the emotion label to the server.

[0981] Step 7:

[0982] The server inputs the determined emotion labels into the mentoring generative model.

[0983] Step 8:

[0984] The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as "Please take an early rest today."

[0985] Step 9:

[0986] The server inputs the generated feedback into a speech synthesis engine, which converts the text-format feedback into speech data.

[0987] Step 10:

[0988] The terminal plays back the converted audio data and provides feedback to the user.

[0989] Step 11:

[0990] The user listens to the audio feedback from the device and takes appropriate action based on it, such as taking an early rest.

[0991] Through this series of steps, users can have their emotions detected from multiple angles using the emotion engine, and receive optimal feedback in real time.

[0992] Example 2

[0993] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0994] Conventional emotion recognition systems detect emotions based solely on the user's voice data, resulting in insufficient emotion recognition accuracy. Furthermore, the quality and applicability of the feedback provided is limited, resulting in insufficient user response. Therefore, there is a need for a system that can improve emotion recognition accuracy and provide more appropriate feedback.

[0995] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0996] In this invention, the server includes means for recognizing emotions from multiple data sources including facial expressions and body movements, means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, and means for transmitting the generated feedback to the user by voice, thereby improving the accuracy of emotion recognition and enabling more appropriate and useful feedback to be provided to the user in real time.

[0997] - "Multiple data sources including facial expressions and body movements" is a general term for input methods such as sensors, cameras, and microphones that collect multiple types of data, such as a user's facial expressions, body movements, and even voice data.

[0998] "Means for recognizing emotions" refers to a system or algorithm that analyzes collected voice data, facial expression data, and body movement data to identify a user's emotions.

[0999] An "emotion detection model" refers to a machine learning model or algorithm that analyzes a user's voice data and detects emotions from its characteristics, such as tone, speed, and volume.

[1000] "Mentoring generative model" means a system or algorithm that generates appropriate feedback and advice for a user based on emotional information obtained from an emotion detection model.

[1001] "Means for generating feedback" refers to software or a program that uses an emotion detection model or a mentoring generation model to generate advice or guidance according to the user's emotions.

[1002] "Means for transmitting the generated feedback to the user by voice" refers to a system that uses a speech synthesis engine to convert text-based feedback into audio format and transmits it to the user through speakers or earphones.

[1003] An embodiment of the present invention is a system that recognizes emotions using a user's voice data, facial expressions, and body movements, and provides appropriate feedback in real time.

[1004] System Overview

[1005] The system mainly consists of the following elements:

[1006] 1. Data sources including facial expressions and body movements

[1007] 2. Means of recognizing emotions (emotion engine)

[1008] 3. Emotion Detection Model

[1009] 4. Mentoring Generative Model

[1010] 5. Speech synthesis engine

[1011] System configuration

[1012] 1. Data Source:

[1013] The user speaks to the device, and the device's built-in camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data.

[1014] As a specific example, consider the case where a user says, "I'm really tired today." At this time, the device captures the user's tired facial expression, slumped shoulders, and other aspects of their posture, and the microphone records their voice.

[1015] 2. Means of Emotion Recognition (Emotion Engine):

[1016] The device sends the collected data to a server in real time. The emotion engine on the server analyzes this multidimensional data and recognizes the user's emotions. Specifically, it uses facial expression recognition algorithms and motion analysis algorithms.

[1017] 3. Emotion detection model:

[1018] The server inputs the voice data into an emotion detection model, which analyzes voice characteristics such as tone, speed, and volume to detect emotions. It also uses past emotion data to improve the accuracy of emotion detection.

[1019] 4. Mentoring generative model:

[1020] The server receives the output of the emotion detection model and generates appropriate feedback using the mentoring generation model. The generated feedback includes advice and comments according to the user's emotional state. For example, the generated feedback might be "Please go to bed early today."

[1021] 5. Speech synthesis engine:

[1022] The server inputs the generated feedback into a speech synthesis engine, converting the text into natural-sounding speech data, which is then sent to the device.

[1023] 6. Providing feedback:

[1024] The device plays back the voice data sent from the speech synthesis engine and provides feedback to the user. For example, a voice message saying "Please go to bed early today" is played through the speaker or earphones.

[1025] Prompt Sentence Examples

[1026] Below are some example prompts to ensure the system works properly:

[1027] Capture the voice data, facial expressions, and body movements of a user saying, "I'm really tired today," and input them into the emotion engine. After that, once the emotion label "fatigue" is generated, reanalyze it with the emotion detection model, and finally generate the advice "Please take an early rest today," and convert it into voice data with the speech synthesis engine.

[1028] This system recognizes the user's emotions with high accuracy from multiple data sources, including not only their voice but also their facial expressions and body movements, and provides real-time feedback according to their current state, making it possible to more appropriately support the user's physical and mental state.

[1029] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1030] Step 1:

[1031] The user speaks to the device, and the device's camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data (input: user's voice, facial expressions, and body movements; output: multiple data sources).

[1032] Step 2:

[1033] The device transmits the captured facial expression data, body movement data, and voice data to the server in real time (input: multiple data sources, output: data transfer to the server).

[1034] Step 3:

[1035] The server inputs the received multidimensional data into the emotion engine, which then integrates the data and identifies the user's emotion using facial expression recognition and motion analysis algorithms (input: data transferred to the server, output: recognized emotion label).

[1036] Step 4:

[1037] The server inputs the emotion label and audio data output from the emotion engine into the emotion detection model, which analyzes features such as tone, speed, and volume of the audio to re-detect the emotion in detail (input: emotion label, audio data, output: confirmed emotion label).

[1038] Step 5:

[1039] The server inputs the determined emotion label into the mentoring generative model and generates feedback. The mentoring generative model generates appropriate advice based on the identified emotion (input: determined emotion label, output: generated feedback).

[1040] Step 6:

[1041] The server inputs the generated feedback into a speech synthesis engine and converts the text feedback into speech data (input: generated feedback text, output: speech data).

[1042] Step 7:

[1043] The device receives the voice data sent from the server and provides feedback to the user through a speaker or earphones (input: voice data, output: voice message to the user).

[1044] Specific operation example

[1045] 1. The user says to the device, "I'm really tired today."

[1046] 2. The device collects the user's facial expressions, body movements, and voice data and sends it to the server.

[1047] 3. The server uses an emotion engine to recognize the emotion "fatigue."

[1048] 4. The server performs detailed analysis of the voice data using an emotion detection model to reconfirm the emotion.

[1049] 5. The server uses a mentoring generation model to generate advice such as "Take a rest early today," and converts it into voice data using a speech synthesis engine.

[1050] 6. The device plays the generated voice data and provides the user with feedback such as "Please go to bed early today."

[1051] (Application example 2)

[1052] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1053] In autonomous vehicles, the challenge is to monitor the emotional state of passengers in real time and improve their comfort and safety. This will provide an environment where passengers can use autonomous vehicles safely and without anxiety or stress, even for long periods of time.

[1054] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1055] In this invention, the server includes means for detecting emotions from the user's voice data, facial expression data, and body movement data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for providing the feedback to the infotainment system of the autonomous vehicle, and means for communicating the generated feedback to the user by voice. This makes it possible to recognize the passenger's emotional state in real time and provide advice or adjust environmental settings (music, lighting, interior temperature) according to the state.

[1056] An "emotion detection model" is a system for detecting emotions by analyzing a user's voice data, facial expression data, and body movement data.

[1057] The "mentoring generation model" is a system that generates appropriate feedback and advice based on emotions detected by the emotion detection model.

[1058] An "infotainment system" is a system that provides information and entertainment functions and is installed in an autonomous vehicle.

[1059] "Feedback" refers to advice and information generated by the emotion detection model and mentoring generation model according to the user's emotional state.

[1060] "Voice data" is a digital representation of a voice signal containing the content of a user's speech.

[1061] "Facial expression data" refers to a facial expression of a user captured by a photographing device such as a camera and expressed in digital form.

[1062] "Body movement data" refers to the user's body movements captured by sensors or other devices and expressed in digital form.

[1063] An "automated vehicle" is a vehicle whose driving operation is performed by an automatic control system.

[1064] This paper describes an embodiment of the present invention. The system analyzes a user's voice data, facial expression data, and body movements to recognize emotions and improve passenger comfort and safety in an autonomous vehicle. Specifically, it consists of the following four main components:

[1065] 1. Emotion Engine

[1066] 2. Emotion Detection Model

[1067] 3. Mentoring Generative Model

[1068] 4. Speech synthesis engine

[1069] Emotion Engine

[1070] The emotion engine captures the user's voice data, facial expression data, and body movement data. When a user gets into an autonomous vehicle, cameras and sensors installed in the vehicle capture this data and send it to a server. This uses tools such as PyAudio to capture voice data and OpenCV to capture facial expressions and body movements. The data sent to the server is analyzed as multidimensional data by the emotion engine to recognize the user's emotions.

[1071] Emotion detection model

[1072] The emotion detection model uses the emotional information provided by the emotion engine to perform detailed analysis of voice characteristics such as tone, speed, volume, facial expressions, and body movements. This allows it to accurately determine the user's emotional state. The model is built using machine learning frameworks such as TensorFlow and Keras. For example, if the user is relaxed, it generates a label such as "calm," and if they are nervous, it generates a label such as "nervous."

[1073] Mentoring Generative Model

[1074] The mentoring generative model generates appropriate feedback based on the emotion labels obtained from the emotion detection model. The server inputs the emotion labels received from the emotion detection model into a generative AI model (e.g., GPT-2) and generates feedback in text format. This feedback may include advice based on the passenger's emotions or adjustments to environmental settings (music, lighting, interior temperature, etc.).

[1075] Text-to-speech engine

[1076] The generated feedback is input into a speech synthesis engine and converted into voice format. The server then transmits this voice data to the passenger through the infotainment system in the autonomous vehicle. Specific examples of feedback could be, "Take a short break today" or "We'll lower the temperature inside the car a little."

[1077] Specific use cases

[1078] For example, if a user in an autonomous vehicle says, "Oh, I'm really tired today," a camera and microphone installed in the infotainment system capture their voice, facial expressions, and body movements. The voice and facial expression data are sent to a server and analyzed by an emotion engine. If the emotion detection model determines that the user is "fatigue," the mentoring generation model generates feedback such as "Please take an early rest today," which is then converted into voice data by a speech synthesis engine and provided to the user. This feedback contributes to creating an environment where the user can feel safe and relaxed.

[1079] Example prompt sentence:

[1080] "The passenger is experiencing fatigue. Please generate appropriate feedback for this condition:"

[1081] This will allow all passengers in self-driving vehicles to enjoy travel in a comfortable and safe environment.

[1082] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1083] Step 1:

[1084] When a user speaks or makes an action inside an autonomous vehicle, the infotainment system's camera and microphone capture the user's voice data, facial expression data, and body movement data. Specifically, the microphone picks up the voice signal, and the camera records the facial expression and body movement data in video format. This data is input to the terminal in real time.

[1085] Step 2:

[1086] The device transmits the captured voice data, facial expression data, and body movement data to the server. This transmission is performed simultaneously with the digital conversion of the data. For example, the voice data is sent to the server in WAV format, the facial expression data is sent to the server in image frames, and the body movement data is sent to the server in motion capture information. At this time, the device encodes the data into the appropriate format.

[1087] Step 3:

[1088] The server inputs the received multidimensional data into the emotion engine, which performs facial recognition and expression analysis using OpenCV, audio analysis using PyAudio, and integrates other sensor data. This engine performs tasks such as generating emotion labels (such as "anger," "sadness," and "joy") from facial expression data, and extracting emotional tone, speed, and volume features from audio data.

[1089] Step 4:

[1090] The server passes the emotion information obtained from the emotion engine to the emotion detection model. The emotion detection model uses machine learning frameworks such as TensorFlow and Keras to determine each emotion. It also references past emotion history data, which reconfirms the user's emotional state and improves prediction accuracy. Finally, a highly accurate emotion label is generated, which the server passes to the next step.

[1091] Step 5:

[1092] The server passes the emotion labels obtained from the emotion detection model to the mentoring generative model. A generative AI model (e.g., GPT-2) is used to generate appropriate feedback based on the emotion labels. Specifically, a message such as "Please take an early rest today" is generated from the emotion label "fatigue." The following prompt sentence is used for this generation: "The passenger is feeling fatigued. Please generate appropriate feedback for this state:" The server then passes the generated text data to the next step.

[1093] Step 6:

[1094] The server inputs the generated feedback into a speech synthesis engine, which uses text-to-speech technology to convert the text data into voice data. The converted voice data is then transmitted to the user through the infotainment system. For example, a voice message such as "Please go to bed early today" is played.

[1095] Step 7:

[1096] The device then provides the generated voice data to the user, specifically, voice feedback played through the infotainment system speaker, allowing the user to receive appropriate advice and reassurance.

[1097] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1098] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1099] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1100] [Fourth embodiment]

[1101] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1102] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1103] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1104] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1105] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1106] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1107] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1108] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1109] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1110] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1111] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1112] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1113] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1114] As an embodiment of the present invention, a system using an emotion detection model and a mentoring generation model will be described. This system analyzes a user's voice data in real time and provides appropriate feedback.

[1115] Overall system overview

[1116] The system consists of three main components:

[1117] 1. Emotion detection model

[1118] 2. Mentoring Generative Model

[1119] 3. Speech synthesis engine

[1120] Emotion detection model

[1121] The emotion detection model identifies emotions from a user's voice data. When a user speaks to a device (such as a smart speaker), the voice data is captured and sent to a server. The server inputs this voice data into the emotion detection model and generates an emotion label. The emotion detection model analyzes features such as tone, speed, and volume of the voice to identify the emotion the user is feeling (e.g., anger, sadness, joy, etc.).

[1122] Mentoring Generative Model

[1123] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. The mentoring generative model generates useful feedback for the user according to the identified emotion. For example, if the user is feeling stressed, it generates advice such as "Take a deep breath and relax."

[1124] Text-to-speech engine

[1125] The generated feedback is returned in text format to the server, which then inputs this text into a speech synthesis engine to convert it into a speech format that is easy for the user to understand. The converted speech is then transmitted to the user via the terminal.

[1126] Specific use cases

[1127] 1. Emotion detection and feedback

[1128] 1. The user speaks into the device.

[1129] For example: "I'm really tired today."

[1130] 2. The device captures this audio and sends it to the server as audio data.

[1131] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[1132] 4. The server inputs the generated emotion labels into the mentoring generative model.

[1133] 5. The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as the message "Please take an early rest today."

[1134] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[1135] 7. The device plays the converted audio to the user.

[1136] For example: "Please go to bed early today."

[1137] 8. Users can take appropriate action based on the feedback.

[1138] In this way, users can use the system to quickly detect their emotions and receive appropriate advice and feedback at that moment, which is expected to serve as a useful tool for maintaining mental health.

[1139] The processing flow will be explained below.

[1140] Step 1:

[1141] The user speaks to the device, for example, saying, "I'm really tired today."

[1142] Step 2:

[1143] The device captures this audio and converts it into digital audio data.

[1144] Step 3:

[1145] The terminal transmits digital audio data to the server.

[1146] Step 4:

[1147] The server inputs the received audio data into the emotion detection model.

[1148] Step 5:

[1149] An emotion detection model analyzes the tone, rate, and volume of the speech and generates an emotion label (e.g., "fatigue").

[1150] Step 6:

[1151] The server inputs the generated emotion labels into the mentoring generative model.

[1152] Step 7:

[1153] Based on the emotion labels, the mentoring generative model generates appropriate feedback such as "Please take a rest early today."

[1154] Step 8:

[1155] The server inputs the generated message into a speech synthesis engine and converts it into speech format.

[1156] Step 9:

[1157] The terminal plays back the converted audio data and provides feedback to the user.

[1158] Step 10:

[1159] The user listens to the audio feedback from the device and takes appropriate action, such as taking an early rest.

[1160] Through this series of steps, users' emotions are detected in real time and they can receive appropriate advice on the spot.

[1161] Example 1

[1162] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1163] Conventional emotion detection systems have the drawback of being unable to provide real-time feedback and promptly provide appropriate advice based on the user's emotional state. Another problem is that multiple methods for improving the accuracy of emotion detection and the quality of feedback are not integrated.

[1164] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1165] In this invention, the server includes means for acquiring and transmitting user voice data to the server, means for the server to receive the voice data, input the voice data to an emotion detection model, and analyze the voice features to generate an emotion label, means for the server to input the emotion label to a mentoring generation model and generate an appropriate feedback message, means for inputting the generated feedback message to a voice synthesis engine and converting it into a voice format, and means for transmitting the converted voice to the terminal and playing it back to the user, thereby enabling real-time emotion detection and immediate feedback provision.

[1166] A "user" is a human entity that provides speech data to the system and receives feedback.

[1167] A "terminal" is a device for capturing and transmitting audio data to a server, and is a device for playing back the feedback audio.

[1168] A "server" is a computing system that receives audio data, processes the data using the emotion detection model and the mentoring generation model, and generates feedback.

[1169] An "emotion detection model" is an artificial intelligence model that analyzes characteristics of voice data, such as tone, speed, and volume, to identify a user's emotions.

[1170] An "emotion label" is data that indicates the emotional state of a user extracted from voice data by an emotion detection model.

[1171] The "Mentoring Generative Model" is an artificial intelligence model that generates appropriate feedback messages based on emotion labels.

[1172] A "feedback message" is a text message containing advice or instructions that correspond to the user's emotional state.

[1173] "Speech synthesis engine" means software or a device that converts textual feedback messages into speech.

[1174] "Converted speech" is speech data generated from a text message by a speech synthesis engine.

[1175] "Real-time" refers to processing or feedback that occurs immediately or with very little delay.

[1176] This invention is a system that uses an emotion detection model, a mentoring generation model, and a speech synthesis engine to analyze a user's voice data in real time and provide appropriate feedback. A specific embodiment of this system is described below.

[1177] Overall system configuration

[1178] The system consists of the following main components:

[1179] 1. Emotion detection model

[1180] 2. Mentoring Generative Model

[1181] 3. Speech synthesis engine

[1182] Emotion detection model

[1183] The user provides voice data to the device. For example, the user might say, "I'm really tired today." The device captures the voice data and sends it to the server. The server inputs the received voice data into an emotion detection model. The emotion detection model analyzes features such as tone, speed, and volume of the voice and generates an emotion label. The specific software used is, for example, LibROSA, which extracts the features of the voice data, and a generative AI model based on TensorFlow or PyTorch.

[1184] Mentoring Generative Model

[1185] The server obtains the emotion labels obtained from the emotion detection model and inputs them into the mentoring generative model. The mentoring generative model generates useful feedback for the user based on the identified emotion. For example, if the user is feeling stressed, the mentoring generative model generates a message such as "Please take an early rest today." Generative AI models such as GPT-3 are used here.

[1186] Text-to-speech engine

[1187] The generated feedback message is returned in text format to the server, which then inputs this text into a speech synthesis engine to convert it into a speech format that is easy for the user to understand. Specific software used includes the Google Text-to-Speech API and Amazon Polly. The converted speech is then sent to the device and played back to the user.

[1188] Specific use cases

[1189] A specific example of the system's operation is shown below.

[1190] 1. The user speaks to the device, saying, "I'm really tired today."

[1191] 2. The device captures this audio and sends it to the server as audio data.

[1192] 3. The server inputs the received audio data into the emotion detection model, which analyzes the tone, rate, and volume of the audio and generates an emotion label of "fatigue."

[1193] 4. The server inputs the generated emotion labels into the mentoring generative model.

[1194] 5. The mentoring generative model generates an appropriate feedback message based on the emotion label "fatigue," such as "Please take an early rest today."

[1195] 6. The server inputs this message into a speech synthesis engine and converts it into speech format.

[1196] 7. The device plays the converted audio to the user, telling them, "Please go to bed early today."

[1197] 8. Users can take appropriate action based on the feedback.

[1198] A concrete example of a prompt for a generative AI model

[1199] 1. Example prompt for emotion detection model:

[1200] "Based on this audio data, identify the user's current emotion."

[1201] 2. Example prompts for the Mentoring Generative Model:

[1202] "When the emotion label is 'fatigue', generate an appropriate feedback message."

[1203] 3. Example prompt for the speech synthesis engine:

[1204] "Please convert the text 'Please go to bed early today' into audio format."

[1205] In this way, users receive rapid, real-time feedback from the system and can take action to improve their physical and mental health.

[1206] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1207] Step 1:

[1208] The user provides voice data to the terminal, which then captures the voice with a microphone and converts it into digital data.

[1209] Input: User's voice

[1210] Output: Digitized audio data

[1211] Specific operations: Records audio with a microphone, converts the audio data into digital format, and stores the digital data in temporary memory.

[1212] Step 2:

[1213] The device prepares the captured audio data to be sent to the server, dividing it into packets and sending them via TCP / IP.

[1214] Input: Digitized voice data

[1215] Output: Audio data packets to the server

[1216] Specific operation: Packetizes digital data into a specific format and sends it over a network to a server.

[1217] Step 3:

[1218] The server analyzes the received audio data packets and reconstructs the audio data, which is then input into the emotion detection model.

[1219] Input: Voice data packets

[1220] Output: Reconstructed speech data, input data for emotion detection model

[1221] Specific operation: Receive packets, reassemble them appropriately, and combine them into a single audio file. Input the audio into the emotion detection model.

[1222] Step 4:

[1223] The server inputs the audio data into an emotion detection model, analyzes audio features (e.g., tone, rate, volume) and generates emotion labels.

[1224] Input: Reconstructed audio data

[1225] Output: Emotion label

[1226] Specific operation: Extract features from audio data using a library such as LibROSA. Input the extracted features into a generative AI model (e.g., TensorFlow or PyTorch) to obtain emotion labels.

[1227] Step 5:

[1228] The server inputs the emotion labels into a mentoring generative model to generate appropriate feedback messages.

[1229] Input: Emotion label

[1230] Output: Feedback message

[1231] Specific action: The emotion label is input as a prompt sentence into the mentoring generative model to obtain a feedback message (e.g., "Please go to bed early today").

[1232] Step 6:

[1233] The server inputs the generated feedback message into a speech synthesis engine and converts it into speech format.

[1234] Input: Feedback message

[1235] Output: Audio data

[1236] Specific behavior: Converts text messages into speech using the Google Text-to-Speech API, Amazon Polly, etc.

[1237] Step 7:

[1238] The server transmits the feedback message converted into audio format to the terminal.

[1239] Input: Audio data

[1240] Output: Audio data packets

[1241] Specific operation: The generated voice data is packetized for transmission to the terminal and sent via TCP / IP.

[1242] Step 8:

[1243] The device analyzes and reconstructs the received audio data and plays it back through the speaker, thereby providing feedback to the user.

[1244] Input: Voice data packets

[1245] Output: Reconstructed speech data, speech feedback to the user

[1246] Specific operation: The received packets are reconstructed and compiled into a single audio file, which is then played back through the speaker to the user.

[1247] (Application example 1)

[1248] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1249] Conventional food delivery systems have difficulty providing appropriate support and meal suggestions based on the user's emotional state, and have been unable to increase the user's psychological satisfaction. Furthermore, when a user feels fatigued or stressed, there is a lack of a way to provide real-time advice tailored to that emotional state. Therefore, there is a need for improved user experience and psychological health support.

[1250] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1251] In this invention, the server includes means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for communicating the generated feedback to the user by voice, means for including meal suggestions as part of the feedback, and means for including an interface for the user to order food. This makes it possible to provide appropriate feedback and advice according to the user's emotional state, and further to suggest meals and order food delivery according to the emotions.

[1252] An "emotion detection model" is an artificial intelligence model that analyzes a user's voice data and identifies emotions from characteristics such as tone, speed, and volume of the voice.

[1253] The "mentoring generation model" is an artificial intelligence model that generates useful feedback and advice for users based on detected emotions.

[1254] "Speech synthesis engine" refers to software that converts the generated textual feedback into speech.

[1255] "Feedback" refers to appropriate advice or guidance provided based on the user's emotional state.

[1256] "Meal suggestion" refers to selecting and suggesting meals suitable for the user based on the user's emotional state.

[1257] An "interface" refers to the operating screen and input means that a user uses to interact with the system and order food.

[1258] "Voice data" refers to the spoken words or sounds of a user and is digital data captured by a voice recognition system.

[1259] An "emotion label" is an indicator or tag that indicates a user's emotion identified by an emotion detection model.

[1260] As an embodiment of the present invention, a system for analyzing user voice data, detecting emotions, and providing appropriate feedback based on the emotions in a food delivery system will be described. This system mainly consists of the following components:

[1261] 1. Emotion detection model

[1262] 2. Mentoring Generative Model

[1263] 3. Speech synthesis engine

[1264] 4. Meal suggestion tools

[1265] 5. Food ordering interface

[1266] Emotion detection model

[1267] The server receives voice data from users speaking into devices such as smartphones. This voice data is input into an emotion detection model, which analyzes features such as tone, speed, and volume of the voice and generates emotion labels, such as "fatigue" or "joy."

[1268] Mentoring Generative Model

[1269] The server generates appropriate feedback using a mentoring generation model based on the emotion labels obtained from the emotion detection model. This feedback includes advice and guidance based on the emotion the user is feeling. For example, if the user has an emotion label of "fatigue," feedback including a specific meal suggestion such as "hot soup would be good" is generated.

[1270] Text-to-speech engine

[1271] The generated feedback is sent from the server to a speech synthesis engine, which converts the feedback text into speech, which is then transmitted to the user via the terminal.

[1272] Meal suggestion tool

[1273] The feedback generated by the mentoring generative model includes specific meal suggestions based on the user's emotional state, which are structured to help the user choose appropriate meals.

[1274] Food Ordering Interface

[1275] An interface is provided for the user to order food based on the suggested meal menu, and the user can easily use this interface to order the suggested meal.

[1276] Specific examples

[1277] For example, suppose a user says to the device, "I'm tired today, so I want to eat something comforting." The device captures the voice data and sends it to the emotion detection model. The server obtains the emotion label "fatigue" and uses the mentoring generative model to generate feedback such as "If you're feeling tired, a warm soup would be good." This feedback is sent to the speech synthesis engine, converted into audio format, and transmitted to the user. The user can easily order the suggested meal through the provided food ordering interface.

[1278] Prompt Sentence Examples

[1279] "User voice text: 'I'm tired today and I want to eat something comforting.'

[1280] Emotion detection model output: 'Fatigue'

[1281] Generate mentoring messages based on this emotion.”

[1282] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1283] Step 1:

[1284] The user speaks to the device and inputs voice data. For example, if the user says, "I'm tired today, so I want to eat something soothing," the voice is captured by the device.

[1285] Step 2:

[1286] The device sends the captured audio data to the server, which then becomes the input data for the system.

[1287] Step 3:

[1288] The server inputs the received voice data into an emotion detection model, which analyzes the tone, speed, volume, etc. of the voice to identify the user's emotion label. In this step, an emotion label such as "fatigue" is output.

[1289] Step 4:

[1290] The server inputs the generated emotion labels into the mentoring generative model, which then generates an appropriate feedback message based on the emotion labels. For example, for the emotion label "fatigue," the model generates the message "If you're feeling tired, warm soup would be good."

[1291] Step 5:

[1292] The server inputs the generated feedback message into a speech synthesis engine, which converts the text message into a voice message. Specifically, the text "If you're feeling tired, a warm soup would be good" is converted into voice data.

[1293] Step 6:

[1294] The server transmits the converted voice data to the terminal, which then plays the voice data and transmits it to the user, so that the user receives voice feedback.

[1295] Step 7:

[1296] The server generates a message containing specific meal suggestions as part of the feedback, such as a specific menu item like "hot soup," in a format that makes it easy for the user to choose.

[1297] Step 8:

[1298] The terminal provides a user interface that displays the suggested meal menu and provides a means for the user to perform operations to order it. The user can use this interface to easily order the suggested meal.

[1299] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1300] As an embodiment of the present invention, we will explain a system that combines an emotion engine that recognizes a user's emotions using information such as facial expressions and body movements in addition to voice data. This system quickly detects a user's diverse emotional information and provides appropriate feedback in real time.

[1301] Overall system overview

[1302] The system consists of four main components:

[1303] 1. Emotion Engine

[1304] 2. Emotion Detection Model

[1305] 3. Mentoring Generative Model

[1306] 4. Speech synthesis engine

[1307] Emotion Engine

[1308] The emotion engine recognizes emotions from the user's voice data, facial expressions, and body movements. When the user is close to the device, the device's camera and sensors capture facial expressions and body movements. In addition, voice data is also acquired. This data is sent to the server and input into the emotion engine. The emotion engine integrates this multidimensional data and recognizes the user's emotions (e.g., joy, anger, sadness, surprise, etc.) with high accuracy.

[1309] Emotion detection model

[1310] The emotion detection model further analyzes the voice data based on the emotional information provided by the emotion engine. This model further analyzes voice characteristics such as tone, speed, and volume to determine the emotion. It can also refer to the user's past emotional history to predict the user's current emotional state.

[1311] Mentoring Generative Model

[1312] The mentoring generative model generates appropriate feedback and advice based on the emotion labels obtained from the emotion detection model. The server obtains the emotion labels from the emotion detection model and inputs them into the mentoring generative model. This model generates advice according to the identified emotion and provides useful feedback to the user.

[1313] Text-to-speech engine

[1314] The generated feedback is returned to the server in text format and converted into audio format by inputting it into a speech synthesis engine. The device plays back this audio data and conveys the feedback to the user.

[1315] Specific use cases

[1316] 1. Recognizing emotions and providing feedback

[1317] 1. As the user speaks to the device, data including facial expressions and body movements is captured by the device's camera and sensors.

[1318] For example: "I'm really tired today."

[1319] 2. The device transmits this voice data, facial expression data, and body movement data to the server.

[1320] 3. The server inputs the received multidimensional data into the emotion engine, which analyzes the data and generates an emotion label, such as "fatigue."

[1321] 4. The server re-inputs the audio data, along with the emotion labels obtained from the emotion engine, into the emotion detection model, which analyzes the audio features and reconfirms the results from the emotion engine.

[1322] 5. The server requests the mentoring model to generate feedback based on the determined emotion label. In response to the emotion "fatigue," the mentoring model generates advice such as "Take an early rest today."

[1323] 6. The server inputs this advice into a speech synthesis engine and converts it into voice data.

[1324] 7. The device plays the generated audio data and provides feedback to the user.

[1325] For example: "Please go to bed early today."

[1326] 8. The user follows the feedback and takes appropriate action.

[1327] This system allows users to receive accurate feedback in real time based on a variety of emotional data, which is expected to provide more effective support for users' mental health.

[1328] The processing flow will be explained below.

[1329] Step 1:

[1330] As a user speaks to the device, the device's camera and sensors capture the user's facial expressions and body movements. For example, the user might say, "I'm really tired today."

[1331] Step 2:

[1332] The device digitizes the captured voice data, facial expression data, and body movement data and transmits them to a server.

[1333] Step 3:

[1334] The server inputs the received voice data, facial expression data, and body movement data into the emotion engine, which integrates and analyzes the data to identify the user's emotional state.

[1335] Step 4:

[1336] Based on the analysis results, the emotion engine generates an emotion label called "fatigue" and returns it to the server.

[1337] Step 5:

[1338] The server inputs the emotion labels and audio data obtained from the emotion engine into the emotion detection model, which analyzes the tone, rate, and volume of the audio and double-checks the results from the emotion engine.

[1339] Step 6:

[1340] The emotion detection model determines the emotion "fatigue" and returns the emotion label to the server.

[1341] Step 7:

[1342] The server inputs the determined emotion labels into the mentoring generative model.

[1343] Step 8:

[1344] The mentoring generative model generates appropriate feedback based on the emotion label "fatigue," such as "Please take an early rest today."

[1345] Step 9:

[1346] The server inputs the generated feedback into a speech synthesis engine, which converts the text-format feedback into speech data.

[1347] Step 10:

[1348] The terminal plays back the converted audio data and provides feedback to the user.

[1349] Step 11:

[1350] The user listens to the audio feedback from the device and takes appropriate action based on it, such as taking an early rest.

[1351] Through this series of steps, users can have their emotions detected from multiple angles using the emotion engine, and receive optimal feedback in real time.

[1352] Example 2

[1353] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1354] Conventional emotion recognition systems detect emotions based solely on the user's voice data, resulting in insufficient emotion recognition accuracy. Furthermore, the quality and applicability of the feedback provided is limited, resulting in insufficient user response. Therefore, there is a need for a system that can improve emotion recognition accuracy and provide more appropriate feedback.

[1355] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1356] In this invention, the server includes means for recognizing emotions from multiple data sources including facial expressions and body movements, means for detecting emotions from user voice data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, and means for transmitting the generated feedback to the user by voice, thereby improving the accuracy of emotion recognition and enabling more appropriate and useful feedback to be provided to the user in real time.

[1357] - "Multiple data sources including facial expressions and body movements" is a general term for input methods such as sensors, cameras, and microphones that collect multiple types of data, such as a user's facial expressions, body movements, and even voice data.

[1358] "Means for recognizing emotions" refers to a system or algorithm that analyzes collected voice data, facial expression data, and body movement data to identify a user's emotions.

[1359] An "emotion detection model" refers to a machine learning model or algorithm that analyzes a user's voice data and detects emotions from its characteristics, such as tone, speed, and volume.

[1360] "Mentoring generative model" means a system or algorithm that generates appropriate feedback and advice for a user based on emotional information obtained from an emotion detection model.

[1361] "Means for generating feedback" refers to software or a program that uses an emotion detection model or a mentoring generation model to generate advice or guidance according to the user's emotions.

[1362] "Means for transmitting the generated feedback to the user by voice" refers to a system that uses a speech synthesis engine to convert text-based feedback into audio format and transmits it to the user through speakers or earphones.

[1363] An embodiment of the present invention is a system that recognizes emotions using a user's voice data, facial expressions, and body movements, and provides appropriate feedback in real time.

[1364] System Overview

[1365] The system mainly consists of the following elements:

[1366] 1. Data sources including facial expressions and body movements

[1367] 2. Means of recognizing emotions (emotion engine)

[1368] 3. Emotion Detection Model

[1369] 4. Mentoring Generative Model

[1370] 5. Speech synthesis engine

[1371] System configuration

[1372] 1. Data Source:

[1373] The user speaks to the device, and the device's built-in camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data.

[1374] As a specific example, consider the case where a user says, "I'm really tired today." At this time, the device captures the user's tired facial expression, slumped shoulders, and other aspects of their posture, and the microphone records their voice.

[1375] 2. Means of Emotion Recognition (Emotion Engine):

[1376] The device sends the collected data to a server in real time. The emotion engine on the server analyzes this multidimensional data and recognizes the user's emotions. Specifically, it uses facial expression recognition algorithms and motion analysis algorithms.

[1377] 3. Emotion detection model:

[1378] The server inputs the voice data into an emotion detection model, which analyzes voice characteristics such as tone, speed, and volume to detect emotions. It also uses past emotion data to improve the accuracy of emotion detection.

[1379] 4. Mentoring generative model:

[1380] The server receives the output of the emotion detection model and generates appropriate feedback using the mentoring generation model. The generated feedback includes advice and comments according to the user's emotional state. For example, the generated feedback might be "Please go to bed early today."

[1381] 5. Speech synthesis engine:

[1382] The server inputs the generated feedback into a speech synthesis engine, converting the text into natural-sounding speech data, which is then sent to the device.

[1383] 6. Providing feedback:

[1384] The device plays back the voice data sent from the speech synthesis engine and provides feedback to the user. For example, a voice message saying "Please go to bed early today" is played through the speaker or earphones.

[1385] Prompt Sentence Examples

[1386] Below are some example prompts to ensure the system works properly:

[1387] Capture the voice data, facial expressions, and body movements of a user saying, "I'm really tired today," and input them into the emotion engine. After that, once the emotion label "fatigue" is generated, reanalyze it with the emotion detection model, and finally generate the advice "Please take an early rest today," and convert it into voice data with the speech synthesis engine.

[1388] This system recognizes the user's emotions with high accuracy from multiple data sources, including not only their voice but also their facial expressions and body movements, and provides real-time feedback according to their current state, making it possible to more appropriately support the user's physical and mental state.

[1389] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1390] Step 1:

[1391] The user speaks to the device, and the device's camera and sensors capture the user's facial expressions and body movements, while the microphone collects audio data (input: user's voice, facial expressions, and body movements; output: multiple data sources).

[1392] Step 2:

[1393] The device transmits the captured facial expression data, body movement data, and voice data to the server in real time (input: multiple data sources, output: data transfer to the server).

[1394] Step 3:

[1395] The server inputs the received multidimensional data into the emotion engine, which then integrates the data and identifies the user's emotion using facial expression recognition and motion analysis algorithms (input: data transferred to the server, output: recognized emotion label).

[1396] Step 4:

[1397] The server inputs the emotion label and audio data output from the emotion engine into the emotion detection model, which analyzes features such as tone, speed, and volume of the audio to re-detect the emotion in detail (input: emotion label, audio data, output: confirmed emotion label).

[1398] Step 5:

[1399] The server inputs the determined emotion label into the mentoring generative model and generates feedback. The mentoring generative model generates appropriate advice based on the identified emotion (input: determined emotion label, output: generated feedback).

[1400] Step 6:

[1401] The server inputs the generated feedback into a speech synthesis engine and converts the text feedback into speech data (input: generated feedback text, output: speech data).

[1402] Step 7:

[1403] The device receives the voice data sent from the server and provides feedback to the user through a speaker or earphones (input: voice data, output: voice message to the user).

[1404] Specific operation example

[1405] 1. The user says to the device, "I'm really tired today."

[1406] 2. The device collects the user's facial expressions, body movements, and voice data and sends it to the server.

[1407] 3. The server uses an emotion engine to recognize the emotion "fatigue."

[1408] 4. The server performs detailed analysis of the voice data using an emotion detection model to reconfirm the emotion.

[1409] 5. The server uses a mentoring generation model to generate advice such as "Take a rest early today," and converts it into voice data using a speech synthesis engine.

[1410] 6. The device plays the generated voice data and provides the user with feedback such as "Please go to bed early today."

[1411] (Application example 2)

[1412] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1413] In autonomous vehicles, the challenge is to monitor the emotional state of passengers in real time and improve their comfort and safety. This will provide an environment where passengers can use autonomous vehicles safely and without anxiety or stress, even for long periods of time.

[1414] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1415] In this invention, the server includes means for detecting emotions from the user's voice data, facial expression data, and body movement data using an emotion detection model, means for generating appropriate feedback based on the detected emotions using a mentoring generation model, means for providing the feedback to the infotainment system of the autonomous vehicle, and means for communicating the generated feedback to the user by voice. This makes it possible to recognize the passenger's emotional state in real time and provide advice or adjust environmental settings (music, lighting, interior temperature) according to the state.

[1416] An "emotion detection model" is a system for detecting emotions by analyzing a user's voice data, facial expression data, and body movement data.

[1417] The "mentoring generation model" is a system that generates appropriate feedback and advice based on emotions detected by the emotion detection model.

[1418] An "infotainment system" is a system that provides information and entertainment functions and is installed in an autonomous vehicle.

[1419] "Feedback" refers to advice and information generated by the emotion detection model and mentoring generation model according to the user's emotional state.

[1420] "Voice data" is a digital representation of a voice signal containing the content of a user's speech.

[1421] "Facial expression data" refers to a facial expression of a user captured by a photographing device such as a camera and expressed in digital form.

[1422] "Body movement data" refers to the user's body movements captured by sensors or other devices and expressed in digital form.

[1423] An "automated vehicle" is a vehicle whose driving operation is performed by an automatic control system.

[1424] This paper describes an embodiment of the present invention. The system analyzes a user's voice data, facial expression data, and body movements to recognize emotions and improve passenger comfort and safety in an autonomous vehicle. Specifically, it consists of the following four main components:

[1425] 1. Emotion Engine

[1426] 2. Emotion Detection Model

[1427] 3. Mentoring Generative Model

[1428] 4. Speech synthesis engine

[1429] Emotion Engine

[1430] The emotion engine captures the user's voice data, facial expression data, and body movement data. When a user gets into an autonomous vehicle, cameras and sensors installed in the vehicle capture this data and send it to a server. This uses tools such as PyAudio to capture voice data and OpenCV to capture facial expressions and body movements. The data sent to the server is analyzed as multidimensional data by the emotion engine to recognize the user's emotions.

[1431] Emotion detection model

[1432] The emotion detection model uses the emotional information provided by the emotion engine to perform detailed analysis of voice characteristics such as tone, speed, volume, facial expressions, and body movements. This allows it to accurately determine the user's emotional state. The model is built using machine learning frameworks such as TensorFlow and Keras. For example, if the user is relaxed, it generates a label such as "calm," and if they are nervous, it generates a label such as "nervous."

[1433] Mentoring Generative Model

[1434] The mentoring generative model generates appropriate feedback based on the emotion labels obtained from the emotion detection model. The server inputs the emotion labels received from the emotion detection model into a generative AI model (e.g., GPT-2) and generates feedback in text format. This feedback may include advice based on the passenger's emotions or adjustments to environmental settings (music, lighting, interior temperature, etc.).

[1435] Text-to-speech engine

[1436] The generated feedback is input into a speech synthesis engine and converted into voice format. The server then transmits this voice data to the passenger through the infotainment system in the autonomous vehicle. Specific examples of feedback could be, "Take a short break today" or "We'll lower the temperature inside the car a little."

[1437] Specific use cases

[1438] For example, if a user in an autonomous vehicle says, "Oh, I'm really tired today," a camera and microphone installed in the infotainment system capture their voice, facial expressions, and body movements. The voice and facial expression data are sent to a server and analyzed by an emotion engine. If the emotion detection model determines that the user is "fatigue," the mentoring generation model generates feedback such as "Please take an early rest today," which is then converted into voice data by a speech synthesis engine and provided to the user. This feedback contributes to creating an environment where the user can feel safe and relaxed.

[1439] Example prompt sentence:

[1440] "The passenger is experiencing fatigue. Please generate appropriate feedback for this condition:"

[1441] This will allow all passengers in self-driving vehicles to enjoy travel in a comfortable and safe environment.

[1442] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1443] Step 1:

[1444] When a user speaks or makes an action inside an autonomous vehicle, the infotainment system's camera and microphone capture the user's voice data, facial expression data, and body movement data. Specifically, the microphone picks up the voice signal, and the camera records the facial expression and body movement data in video format. This data is input to the terminal in real time.

[1445] Step 2:

[1446] The device transmits the captured voice data, facial expression data, and body movement data to the server. This transmission is performed simultaneously with the digital conversion of the data. For example, the voice data is sent to the server in WAV format, the facial expression data is sent to the server in image frames, and the body movement data is sent to the server in motion capture information. At this time, the device encodes the data into the appropriate format.

[1447] Step 3:

[1448] The server inputs the received multidimensional data into the emotion engine, which performs facial recognition and expression analysis using OpenCV, audio analysis using PyAudio, and integrates other sensor data. This engine performs tasks such as generating emotion labels (such as "anger," "sadness," and "joy") from facial expression data, and extracting emotional tone, speed, and volume features from audio data.

[1449] Step 4:

[1450] The server passes the emotion information obtained from the emotion engine to the emotion detection model. The emotion detection model uses machine learning frameworks such as TensorFlow and Keras to determine each emotion. It also references past emotion history data, which reconfirms the user's emotional state and improves prediction accuracy. Finally, a highly accurate emotion label is generated, which the server passes to the next step.

[1451] Step 5:

[1452] The server passes the emotion labels obtained from the emotion detection model to the mentoring generative model. A generative AI model (e.g., GPT-2) is used to generate appropriate feedback based on the emotion labels. Specifically, a message such as "Please take an early rest today" is generated from the emotion label "fatigue." The following prompt sentence is used for this generation: "The passenger is feeling fatigued. Please generate appropriate feedback for this state:" The server then passes the generated text data to the next step.

[1453] Step 6:

[1454] The server inputs the generated feedback into a speech synthesis engine, which uses text-to-speech technology to convert the text data into voice data. The converted voice data is then transmitted to the user through the infotainment system. For example, a voice message such as "Please go to bed early today" is played.

[1455] Step 7:

[1456] The device then provides the generated voice data to the user, specifically, voice feedback played through the infotainment system speaker, allowing the user to receive appropriate advice and reassurance.

[1457] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1458] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1459] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1460] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1461] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1462] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1463] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1464] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1465] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1466] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1467] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1468] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1469] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1470] 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.

[1471] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1472] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1473] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1474] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1475] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1476] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1477] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1478] The following is further disclosed regarding the above embodiment.

[1479] (Claim 1)

[1480] means for detecting emotions from user speech data using an emotion detection model;

[1481] means for generating appropriate feedback based on the detected emotions using a mentoring generative model;

[1482] means for audibly communicating the generated feedback to the user;

[1483] A system including:

[1484] (Claim 2)

[1485] 10. The system of claim 1, wherein the emotion detection model detects emotions by analyzing tone, rate, and volume of speech.

[1486] (Claim 3)

[1487] 10. The system of claim 1, wherein the mentoring generative model generates specific advice based on the detected emotions.

[1488] (Claim 4)

[1489] 10. The system of claim 1, further comprising means for receiving the generated feedback in text form and converting it into speech using a speech synthesis engine.

[1490] (Claim 5)

[1491] 10. The system of claim 1, further comprising communication means for real-time emotion detection and feedback between a user and the system.

[1492] "Example 1"

[1493] (Claim 1)

[1494] means for acquiring user voice data and transmitting it to a server;

[1495] a server receiving the voice data, inputting the voice data into an emotion detection model, and analyzing voice features to generate emotion labels;

[1496] a means for the server to input the emotion label into a mentoring generative model and generate an appropriate feedback message;

[1497] a means for inputting the generated feedback message into a speech synthesis engine and converting it into speech format;

[1498] means for transmitting the converted audio to the terminal and playing it to the user;

[1499] A system including:

[1500] (Claim 2)

[1501] 10. The system of claim 1, wherein the emotion detection model detects emotions by analyzing tone, rate, and volume of speech.

[1502] (Claim 3)

[1503] 10. The system of claim 1, wherein the mentoring generative model generates specific advice based on the detected emotions.

[1504] "Application Example 1"

[1505] (Claim 1)

[1506] means for detecting emotions from user speech data using an emotion detection model;

[1507] means for generating appropriate feedback based on the detected emotions using a mentoring generative model;

[1508] means for audibly communicating the generated feedback to the user;

[1509] measures that include dietary suggestions as part of the feedback;

[1510] means including an interface for a user to order food;

[1511] A system including:

[1512] (Claim 2)

[1513] 10. The system of claim 1, wherein the emotion detection model detects emotions by analyzing tone, rate, and volume of speech.

[1514] (Claim 3)

[1515] 10. The system of claim 1, wherein the mentoring generative model generates specific advice based on the detected emotions.

[1516] "Example 2: Combining Emotion Engines"

[1517] (Claim 1)

[1518] means for recognizing emotions from multiple data sources including facial expressions and body movements;

[1519] means for detecting emotions from user speech data using an emotion detection model;

[1520] means for generating appropriate feedback based on the detected emotions using a mentoring generative model;

[1521] means for audibly communicating the generated feedback to the user;

[1522] A system including:

[1523] (Claim 2)

[1524] 10. The system of claim 1, wherein the emotion detection model detects emotions by analyzing tone, rate, and volume of speech.

[1525] (Claim 3)

[1526] 10. The system of claim 1, wherein the mentoring generative model generates specific advice based on the detected emotions.

[1527] "Application example 2 when combining emotion engines"

[1528] (Claim 1)

[1529] means for detecting emotions from voice data, facial expression data, and body movement data of a user using an emotion detection model;

[1530] means for generating appropriate feedback based on the detected emotions using a mentoring generative model;

[1531] a means for providing feedback to an infotainment system of the autonomous vehicle;

[1532] means for audibly communicating the generated feedback to the user;

[1533] A system including:

[1534] (Claim 2)

[1535] 10. The system of claim 1, wherein the emotion detection model detects emotions by analyzing tone, speed, volume, facial expressions, and body movements of speech.

[1536] (Claim 3)

[1537] 10. The system of claim 1, wherein the mentoring generative model generates specific advice or adjustments to environmental settings (music, lighting, interior temperature) based on the detected emotions. [Explanation of symbols]

[1538] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. means for detecting emotions from user speech data using an emotion detection model; means for generating appropriate feedback based on the detected emotions using a mentoring generative model; means for audibly communicating the generated feedback to the user; A system including:

2. The system of claim 1 , wherein the emotion detection model detects emotions by analyzing tone, rate, and volume of speech.

3. The system of claim 1 , wherein the mentoring generative model generates specific advice based on the detected emotions.

4. 2. The system of claim 1, further comprising means for receiving the generated feedback in text form and converting it into speech using a speech synthesis engine.

5. 10. The system of claim 1, further comprising communication means for real-time emotion detection and feedback between a user and the system.

Images