System

Abstract

A system is provided.SOLUTION: A system, comprising: means for collecting audio; means for pre-processing the collected audio; means for converting the pre-processed audio to text using a AI recognition model; means for translating the converted text using a multilingual generative AI model; means for displaying and playing the translated text; means for expert review and completion of the translated text; means for data encryption; and means for training the generative speech model to account for cultural differences.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 medical settings, accurately assessing the symptoms of elderly or foreign patients can be difficult and time-consuming. Patients also have difficulty understanding the diagnosis and explanations provided by medical professionals, which can delay appropriate medical treatment. This problem is particularly pronounced in multilingual environments and cultural differences, making rapid and accurate communication essential. The present invention aims to solve these problems and enable smooth and effective communication between patients and medical professionals. [Means for solving the problem]

[0005] The present invention is a system including the following means: means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a speech recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for an expert to check and complete the accuracy of the translated text, means for encrypting data, and means for training the generative AI model to take cultural differences into account. This enables smooth and prompt communication between patients and medical professionals, enabling appropriate medical care to be provided.

[0006] "Voice data" refers to information recorded as electrical signals from the speech of a patient or medical professional captured from a terminal.

[0007] "Preprocessing" is the process of performing processes such as noise reduction and normalization on audio data to convert it into a format that is easy to analyze.

[0008] A "speech recognition model" is a machine learning algorithm or model for analyzing speech data and converting it into text data.

[0009] "Text data" is information in the form of a string of characters converted from voice data by a voice recognition model.

[0010] A "multilingual generative AI model" is a generative artificial intelligence model that translates input text data into different languages.

[0011] "Translation" is the process of converting text written in one language into another language.

[0012] "Display and playback" refers to the act of visually showing the translated text on a monitor and outputting it as audio using speech synthesis.

[0013] "Experts" are medical professionals such as doctors and nurses who have a high level of knowledge and experience in the medical field.

[0014] "Checking and completing" is the process by which experts check the accuracy of the translated text and provide corrections or additional information where necessary.

[0015] "Encryption" is a security technology that converts data into a format that cannot be deciphered by third parties.

[0016] "Cultural training" is a learning process that improves a generative AI model's ability to understand and appropriately translate different cultural backgrounds and culturally specific expressions. [Brief explanation of the drawings]

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

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

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

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

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

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

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

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

[0025] [First embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0038] This is a system to support smooth communication between elderly and foreign patients and medical professionals. The system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment.

[0039] Program implementation and processing explanation

[0040] Basic configuration

[0041] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[0042] Overview of the main measures that will emerge

[0043] 1. Collection of audio data

[0044] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[0045] 2. Preprocessing of audio data

[0046] Server: Performs noise reduction and normalization on the received audio data, removing unnecessary noise and converting the audio data into a format that is easier to analyze.

[0047] 3. Speech to text conversion

[0048] Server: The preprocessed speech data is analyzed using a speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[0049] 4. Text Translation

[0050] Server: The converted text data is translated in real time using a multilingual generative AI model. For example, the Japanese text "I have a headache" is translated into English as "I have a headache."

[0051] 5. Display and playback of translated text and audio

[0052] On your device: The translated text will be displayed on your device screen and played back using speech synthesis, if available. For example, the translation will be displayed on your device and the speech synthesis will say "I have a headache."

[0053] 6. Expert check and completion

[0054] User (medical professional): Checks the translation results and corrects or completes them as necessary. This improves the accuracy and reliability of the translation. For example, a medical professional corrects "windpipe" to "trachea."

[0055] 7. Data Encryption and Protection

[0056] Server: All data is protected with advanced encryption technology, for example, using the AES encryption algorithm, and data in transit is sent and received over secure protocols (TLS / SSL).

[0057] 8. Training a generative AI model

[0058] Server: Generative AI models are trained to understand cultural differences and sensitivities, for example, using datasets specific to Japanese customs and expressions.

[0059] Specific examples

[0060] Consider a scenario in which a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital.

[0061] 1. Voice input: The patient speaks to the terminal in English, "I have a headache."

[0062] 2. Data preprocessing and speech recognition: This speech data is sent to the server, where it is noise-reduced and converted to text using an ASR model.

[0063] 3. Text translation: A generative AI model translates the English phrase "I have a headache" into Japanese phrase "I have a headache."

[0064] 4. Display and audio output: The translated Japanese text is displayed on the doctor's device and played as audio if necessary.

[0065] 5. Expert review: A doctor reviews the translation and asks additional questions or makes a diagnosis.

[0066] The above steps will enable rapid and accurate medical communication that transcends language barriers.

[0067] The processing flow will be explained below.

[0068] Step 1:

[0069] Audio data collection

[0070] Device: The patient or healthcare professional speaks into the device. The device's built-in microphone captures the audio and begins recording.

[0071] Example of operation: When a patient says "I have a headache" to the device, the voice is recorded.

[0072] Step 2:

[0073] Sending audio data

[0074] Terminal: Streams recorded audio data to the server in real time.

[0075] How it works: Captured audio data is converted into digital form and sent over a network to a server.

[0076] Step 3:

[0077] Audio data preprocessing

[0078] Server: Performs noise reduction and normalization on the received audio data, converting it into a clear and consistent format.

[0079] How it works: A filtering algorithm is applied to remove background noise and even out volume variations.

[0080] Step 4:

[0081] Text conversion using speech recognition models

[0082] Server: The preprocessed speech data is input into the speech recognition model and converted into text.

[0083] Example of how it works: A speech recognition algorithm analyzes the characteristics of a voice and converts the speech "I have a headache" into the text "I have a headache."

[0084] Step 5:

[0085] Text translation

[0086] Server: Using a generative AI model, translate text data into the target language. Use a multilingual translation engine.

[0087] Example of how it works: The text "I have a headache" is translated into English as "I have a headache".

[0088] Step 6:

[0089] Display and playback of translation results

[0090] On Device: Displays translated text and audio on the user's device, optionally using speech synthesis to play the translation aloud.

[0091] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache" in English.

[0092] Step 7:

[0093] Checked and supplemented by experts

[0094] User (medical professional): Review the translated content and provide corrections or additional information as needed.

[0095] Example of how it works: A doctor checks the translation results and corrects any inaccuracies. For example, he or she asks an additional question in Japanese: "Please tell me the cause of my headache."

[0096] Step 8:

[0097] Data encryption and protection

[0098] Server: We encrypt all data sent and received and use advanced security protocols to prevent unauthorized access to your data by third parties.

[0099] How it works: The server encrypts data using the AES encryption algorithm, and the encrypted data is sent and received via the TLS protocol.

[0100] Step 9:

[0101] Training a generative AI model

[0102] Server: Trains AI models to understand cultural differences and multilingual environments, thereby reducing mistranslations and cultural misunderstandings.

[0103] How it works: An AI model trains on a server using a multilingual dataset to improve its ability to understand Japanese-specific nuances.

[0104] Example 1

[0105] 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."

[0106] In today's medical settings, communication barriers between elderly or foreign patients and medical professionals are a serious problem. It is particularly difficult to communicate smoothly with patients who have language differences, hearing impairments, or cultural backgrounds, which can delay appropriate diagnoses and treatment. A system that can resolve these issues and support accurate and prompt communication in medical settings is needed.

[0107] 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.

[0108] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for experts to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for noise reduction and normalization, means for inputting voice into a voice recognition model, and means for synthesizing voice, thereby enabling accurate and prompt communication between patients and medical professionals with different languages, hearing impairments, or cultural backgrounds.

[0109] An "audio data collection means" is a device or method for capturing and transmitting audio from a patient or medical professional to a server.

[0110] The "audio data preprocessing means" refers to a device or method for converting collected audio data into a form that is easy to analyze by performing processes such as noise reduction and normalization.

[0111] A "speech recognition model" is a machine learning algorithm that analyzes preprocessed speech data and converts it into text data.

[0112] A "multilingual generative AI model" is an artificial intelligence model used to translate between different languages ​​in real time.

[0113] "Display and playback means" refers to a device or method for displaying the translated text on the screen of the terminal and, if necessary, synthesizing it to speech and playing it back.

[0114] "Expert checking and completion means" refers to an interface and method for medical professionals to check the translation results and make corrections or completions as necessary.

[0115] A "data encryption means" is a device or method for encrypting data using an encryption algorithm to protect all data processed within the system.

[0116] A "means for training a generative AI model" is a device or method for training a generative AI model with a particular dataset so that the model understands cultural differences and sensitivities.

[0117] "Noise reduction and normalization means" refers to a device or method for reducing noise in audio data during pre-processing and adjusting the volume and pitch to a consistent level.

[0118] A "means for inputting speech into a speech recognition model" is a device or method for appropriately inputting preprocessed speech data into a speech recognition model.

[0119] "Speech synthesis means" refers to a device or method for generating speech based on text data and reproducing it.

[0120] This invention relates to a system that supports smooth communication between elderly and foreign patients and medical professionals. Specifically, it enables accurate and prompt medical treatment by providing voice, text, translation, and real-time medical information.

[0121] Fundamental System Configuration

[0122] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[0123] Technical Details

[0124] Hardware and Basic Operation

[0125] The device is used by the patient or a medical professional and collects voice data through a built-in microphone. The device can be a smartphone, tablet, or dedicated voice collection device. The collected voice data is sent to a server in real time. For example, if a patient says, "I have a headache," the voice is captured instantly.

[0126] The server preprocesses the received audio data. The main processes are noise reduction and normalization. Noise reduction removes background noise, and normalization maintains the quality of the audio data.

[0127] The preprocessed speech data is then converted into text data using an Automatic Speech Recognition (ASR) model. For example, the Japanese phrase "I have a headache" is converted directly into text.

[0128] The converted text is then translated in real time by a multilingual generative AI model, pre-trained to understand cultural differences and sensitivities, for example, translating the Japanese text "Watashi wa tou ga kaiyuu" (I have a headache) into English.

[0129] The translated text and audio are then displayed and played back on the device. The translated text is displayed on the device screen and the translation is played back using speech synthesis. For example, the translated "I have a headache" is displayed and played back as audio.

[0130] The user (medical professional) can check the translation results through the interface provided by the system and make corrections or additions as necessary, thereby improving the accuracy and reliability of the translation. For example, a medical professional can correct "windpipe" to "trachea."

[0131] The server protects all data with advanced encryption technology, specifically using the AES encryption algorithm, and data in transit is sent and received via secure protocols (TLS / SSL).

[0132] Specific examples

[0133] Consider a case where a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital. The following steps are taken:

[0134] 1. Voice input:

[0135] The patient speaks to the terminal in English, "I have a headache."

[0136] 2. Data preprocessing and speech recognition:

[0137] This audio data is sent to a server, where it is noise reduced and normalized before being converted to text using an ASR model.

[0138] 3. Text Translation:

[0139] The server uses a generative AI model to translate the English phrase "I have a headache" into Japanese phrase "I have a headache."

[0140] 4. Display and audio output:

[0141] The translated Japanese text is displayed on the doctor's device and the audio is played back as "I have a headache."

[0142] 5. Expert review:

[0143] The doctor will check the translated content and provide appropriate diagnosis and treatment.

[0144] Prompt Sentence Examples

[0145] Examples of prompts to enter into this system are:

[0146] "If a patient says, 'I have a headache,' in English, convert that audio data into Japanese text and speech for the doctor to review."

[0147] The above procedures and system configuration will enable rapid and accurate medical communication that transcends language barriers.

[0148] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0149] Step 1:

[0150] Audio data collection

[0151] The device captures the voice of the patient or medical professional. When the patient speaks into the device, the built-in microphone collects the voice and transmits it to the server in real time. The input is voice data, and the output is raw voice data sent to the server.

[0152] Specific behavior:

[0153] The patient speaks into the terminal's microphone, "I have a headache."

[0154] The device collects the voice data and sends it to the server.

[0155] Step 2:

[0156] Audio data preprocessing

[0157] The server performs noise reduction and normalization on the received audio data, converting it into a form that is easier to analyze. The input is raw audio data, and the output is preprocessed audio data.

[0158] Specific behavior:

[0159] The server receives the audio data.

[0160] Apply a noise reduction algorithm to remove background noise.

[0161] Normalizes the volume and pitch of audio data.

[0162] Step 3:

[0163] Speech-to-text conversion

[0164] The server inputs the preprocessed speech data into an Automatic Speech Recognition (ASR) model and converts it into text data. The input is the preprocessed speech data, and the output is text data.

[0165] Specific behavior:

[0166] The server inputs the speech data into the ASR model.

[0167] The ASR model analyzes the speech data and generates the text "I have a headache."

[0168] Step 4:

[0169] Text translation

[0170] The server translates the converted text using a generative AI model. The input is text data, and the output is translated text data.

[0171] Specific behavior:

[0172] The server inputs text into the generative AI model.

[0173] A generative AI model translates the Japanese text "I have a headache" into English "I have a headache."

[0174] Step 5:

[0175] View and play translated text and audio

[0176] The terminal displays the translated text received from the server on the screen and plays it back using speech synthesis if necessary. The input is the translated text data, and the output is the displayed text and speech data.

[0177] Specific behavior:

[0178] The terminal receives the translation result from the server.

[0179] The translated text is displayed on the screen.

[0180] Uses a speech synthesis engine to play the translated text aloud.

[0181] Step 6:

[0182] Checked and supplemented by experts

[0183] The user (medical professional) checks the translation results displayed on the terminal and makes corrections or completions as necessary. The input is the translated text data, and the output is the corrected or completed text data.

[0184] Specific behavior:

[0185] The medical professional checks the translation results displayed on the device's display.

[0186] If the translation is not accurate, correct it.

[0187] Step 7:

[0188] Data encryption and protection

[0189] The server encrypts all audio and text data. The input is the processed data and the output is the encrypted data.

[0190] Specific behavior:

[0191] The server encrypts the audio and text data using the AES encryption algorithm.

[0192] Encrypted data is sent and received using secure protocols (TLS / SSL).

[0193] Step 8:

[0194] Training a generative AI model

[0195] The server trains the generative AI model to understand cultural differences and sensitivities: the input is a specific culture and language dataset, and the output is an optimized generative AI model.

[0196] Specific behavior:

[0197] The server inputs specific culture and language datasets into the AI ​​model.

[0198] The server optimizes the parameters of the AI ​​model and performs training.

[0199] (Application example 1)

[0200] 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."

[0201] To ensure smooth communication between elderly and foreign patients and medical professionals, accurate real-time speech recognition and translation, as well as secure processing and transmission of the data, are necessary. However, current systems lack the ability to process voice data in real time, securely encrypt the data, and train generative AI models that take cultural differences into account, hindering accurate and prompt medical responses.

[0202] 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.

[0203] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a speech recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for an expert to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for securely storing and transmitting the encrypted data, and means for processing the voice data collected by the terminal in real time, thereby enabling real-time and accurate medical communication between patients and medical professionals.

[0204] "Audio data" refers to data in which an audio signal is recorded as digital information.

[0205] A "collection means" is a device or system for acquiring and transmitting audio data to another processing unit.

[0206] The "preprocessing means" is a device or system that performs preprocessing such as noise removal and normalization on collected voice data.

[0207] A "voice recognition model" is an algorithm or AI model that analyzes voice data and converts it into a string of characters (text).

[0208] "Text" is digital data expressed as character information.

[0209] A "multilingual generative AI model" is an AI model that supports multiple languages ​​and translates from one language to another in real time.

[0210] A "translation tool" is a device or system that converts text in a particular language into another language.

[0211] "Display means" refers to a device or interface for visually showing the translation results to the user.

[0212] "Playback means" refers to a device or system that provides the translation result to the user as audio.

[0213] "Means for expert checking and correction" refers to a device or system that allows an expert to verify the accuracy of the translation results and correct them if necessary.

[0214] An "encryption means" is a device or system that encrypts data to protect the security of the data.

[0215] A "training means for a generative AI model to take cultural differences into account" is a device or system that trains a generative AI model to understand differences in cultural background and expression.

[0216] A "real-time processing means" is a device or system that instantly analyzes and processes audio data and provides results to the user in real time.

[0217] "Storage and transmission means" means a device or system that securely stores encrypted data and transmits it wherever needed.

[0218] To implement this invention, a system must be constructed by combining a series of hardware and software. The specific system configuration, processing procedures, and usage methods are described below.

[0219] System configuration

[0220] This system is a multi-function system for speech recognition, real-time translation, data encryption, and displaying and playing back the results.

[0221] Hardware used

[0222] 1. Collection device: A device for collecting voice data, such as a smartphone or head-mounted display. This device has a built-in highly sensitive microphone.

[0223] 2. Servers: High-performance servers are used for data pre-processing, speech recognition, text conversion, translation, and data encryption, including cloud servers.

[0224] 3. Display device: A device that displays the translated data and plays it back as audio. This includes smartphones and tablets.

[0225] Software used

[0226] 1. Speech recognition library: A library for converting voice data to text (e.g., speech_recognition).

[0227] 2. Translation library: A library for translating text into another language (e.g. googletrans).

[0228] 3. Cryptography library: A library for performing data encryption (e.g., the Fernet module for cryptography).

[0229] System processing explanation in natural language

[0230] The server receives voice data from a collection terminal, which is used by a patient or a medical professional to capture the voice data. The captured voice data undergoes pre-processing for noise reduction and normalization before being sent to the server.

[0231] The server converts the preprocessed speech data into text using a speech recognition library. This text data is translated into the specified language in real time using a multilingual generative AI model. For example, if a patient says "I have a headache" in English, this will be translated into Japanese as "I have a headache."

[0232] The translated text is sent to a display terminal where it is displayed on the screen and played back as audio. The system also includes an interface that allows experts to check the accuracy of the translation and make corrections or additions if necessary.

[0233] All data is encrypted using advanced encryption technology and transmitted over secure protocols. The server uses the AES encryption algorithm and TLS / SSL protocols to ensure data security.

[0234] Examples of concrete examples and prompts

[0235] Specific usage scenarios include the following:

[0236] 1. Example: When an English-speaking patient visits a Japanese hospital, they use their smartphone to communicate their symptoms to the doctor. When they speak into their smartphone, for example, "I have a headache," this voice data is translated into Japanese and displayed and played on the doctor's device.

[0237] 2. Example prompt:

[0238] I'm an English-speaking patient with a headache. I say "I have a headache" in English to my smartphone. Please translate the voice data into Japanese and display it so that my doctor can see it. Also, please encrypt the data to ensure its security.

[0239] This invention enables real-time and accurate medical communication between elderly and foreign patients and medical professionals.

[0240] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0241] Step 1:

[0242] The collection terminal captures the speech data uttered by the patient or medical professional. The input is the speech signal collected from the microphone of the collection terminal. This speech signal is converted into a digital format and subjected to noise reduction and normalization processing. The output is the pre-processed speech data.

[0243] Step 2:

[0244] The preprocessed voice data is sent from the collection terminal to the server. The input is the preprocessed voice data. The collection terminal sends the data to the server using a transmission protocol. The output is the voice data received at the server side.

[0245] Step 3:

[0246] The server performs speech recognition processing on the received audio data. The input is noise-reduced and normalized audio data. The server uses a speech recognition library (e.g., speech_recognition) to convert the audio data into text data. The output is text data.

[0247] Step 4:

[0248] The server translates the converted text data using a multilingual generative AI model. The input is text data generated by a speech recognition model. A translation library (e.g., GoogleTrans) is used to translate the text into the specified language in real time. The output is the translated text data.

[0249] Step 5:

[0250] The server encrypts the translated text data. The input is the translated text data. The data is encrypted using an encryption library (e.g. Fernet). The output is the encrypted text data.

[0251] Step 6:

[0252] Encrypted text data is sent from the server to the display terminal. The input is the encrypted text data. The server transfers the data using a secure protocol (e.g., TLS / SSL). The output is the encrypted text data received by the display terminal.

[0253] Step 7:

[0254] The display terminal decrypts encrypted text data received from the server. The input is the encrypted text data. The display terminal uses the encryption library to decrypt the data. The output is the decrypted text data.

[0255] Step 8:

[0256] The display terminal displays and plays back the decoded text data. The input is the decoded text data. The user interface displays the text data on the screen and uses speech synthesis to play the text aloud. The output is a translation result that can be recognized visually and audibly by the user.

[0257] Step 9:

[0258] The expert user uses a display terminal to check the accuracy of the translated text data and make corrections or completions as necessary. The input is the decoded and displayed text data. The user makes corrections through the interface, and the results are reflected in the system. The output is the corrected and completed accurate translation data.

[0259] This series of steps enables accurate and secure medical communication in real time.

[0260] 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.

[0261] This invention is a system for supporting smooth communication and emotion recognition between elderly and foreign patients and medical professionals. This system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it is possible to make diagnoses and treatments that take into account the patient's emotional state.

[0262] Program implementation and processing explanation

[0263] Basic configuration

[0264] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, and a terminal that displays and plays back the converted and translated information. An interface is also provided for experts to check and complete the translation results and recognized emotions. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[0265] Overview of the main measures that will emerge

[0266] 1. Collection of audio data

[0267] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[0268] 2. Sending audio data

[0269] Terminal: Streams recorded audio data to the server in real time. The captured audio data is converted into a digital format and sent over the network to the server.

[0270] 3. Preprocessing of audio data

[0271] Server: Noise reduction and normalization processes are performed on the received audio data, converting the audio into a clear and consistent format. Filtering algorithms are applied to remove background noise and even out volume variations.

[0272] 4. Emotion Recognition by Emotion Engine

[0273] Server: Analyzes the preprocessed audio and video data to recognize the user's emotional state. For example, it identifies emotions such as "anger," "sadness," and "surprise." The emotion recognition algorithm reads the user's emotions from their voice intonation and facial expressions.

[0274] 5. Speech to Text Conversion

[0275] Server: The preprocessed speech data is input into the speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[0276] 6. Text Translation

[0277] Server: Translates text data into the target language using a generative AI model. It uses a multilingual translation engine. For example, the text "I have a headache" is translated into English as "I have a headache."

[0278] 7. Display and playback of translation results and emotions

[0279] Terminal: The translated text and recognized emotion information are displayed on the user's terminal. If necessary, the translation result is played back aloud using speech synthesis. For example, the translation result is displayed on the doctor's terminal, and the speech synthesis reproduces "I have a headache" in English. The display screen also indicates the patient's emotional state (e.g., "anxiety").

[0280] 8. Expert check and completion

[0281] User (medical professional): Review the translated content and the perceived sentiment, and provide corrections or additional information as needed. This improves the accuracy and reliability of the translation. For example, a doctor may review the translation and ask a follow-up question such as "What is causing your headache?", as well as provide suggestions to ease the patient's anxiety.

[0282] 9. Data Encryption and Protection

[0283] Server: All data is protected with advanced encryption technology. Voice and text data is encrypted with the AES encryption algorithm, and data in transit is also securely transmitted via the TLS protocol.

[0284] 10. Training a generative AI model

[0285] Server: Trains AI models to understand cultural differences and multilingual environments. Using multilingual datasets, AI models learn and improve their ability to understand Japanese-specific nuances.

[0286] These steps not only enable fast and accurate medical communication across language barriers, but also provide a more humane response that takes into account the patient's emotional state.

[0287] The processing flow will be explained below.

[0288] Step 1:

[0289] Audio data collection

[0290] Device: The patient or healthcare professional speaks into the device, which has a built-in microphone that captures the voice and records the data in real time.

[0291] Example of operation: When a patient says "I have a headache" to the device, the voice is instantly recorded and saved as digital audio data.

[0292] Step 2:

[0293] Sending audio data

[0294] Terminal: Recorded audio data is sent to the server in streaming format.

[0295] How it works: Send captured audio data to a server using any Wi-Fi or 4G network within range.

[0296] Step 3:

[0297] Audio data preprocessing

[0298] Server: Performs noise reduction and normalization on the received audio data, making it easier to analyze.

[0299] How it works: Using a noise filtering algorithm on the server side to remove background noise and enhance the audio signal.

[0300] Step 4:

[0301] Emotion recognition by emotion engine

[0302] Server: Analyzes the pre-processed audio and video data (if necessary) and recognizes the user's emotions. An emotion recognition algorithm is used.

[0303] Example of how it works: The server analyzes the intonation, speed, and emphasized parts of the voice to identify emotions such as "anger," "sadness," and "surprise." It also uses facial expression data recognized using a camera.

[0304] Step 5:

[0305] Speech-to-text conversion

[0306] Server: The preprocessed speech data is input into the speech recognition model (ASR) and converted into corresponding text data.

[0307] Example of how it works: A speech recognition algorithm listens for speech characteristics and converts the speech "I have a headache" into text "I have a headache."

[0308] Step 6:

[0309] Text translation

[0310] Server: Using a generative AI model, the converted text data is translated into the required language using a multilingual translation engine.

[0311] Working example: Translate the Japanese phrase "I have a headache" into English.

[0312] Step 7:

[0313] Display and playback of translation results and emotional information

[0314] Device: The translated text and recognized emotion information are displayed on the user's device, and the translation result is played aloud using speech synthesis if necessary.

[0315] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache." Emotional information such as "Emotion: Anxiety" is also displayed on the display screen.

[0316] Step 8:

[0317] Checked and supplemented by experts

[0318] User (medical professional): Checks the content of the translated text and sentiment information, and makes corrections or asks additional questions as needed, thereby improving the accuracy of the diagnosis.

[0319] Example of how it works: A doctor checks the translation results and emotional information, and then asks an additional question in Japanese, such as "Please tell me the cause of my headache." The system translates this into English in real time.

[0320] Step 9:

[0321] Data encryption

[0322] Server: All data sent and received is protected with the AES encryption algorithm, and communication is carried out via the TLS / SSL protocol.

[0323] Example of operation: All data, including recorded voice data, translated text, and emotional information, is encrypted and transmitted over a secure communication channel.

[0324] Step 10:

[0325] Training a generative AI model

[0326] Server: Trains generative AI models to understand cultural differences and multilingual environments, using multilingual datasets for training.

[0327] How it works: The server trains an AI model using a dataset on Japanese culture to understand Japanese-specific expressions and nuances.

[0328] The above processing flow enables accurate and prompt communication between patients and medical professionals, and also enables responses that take into account the patient's emotional state.

[0329] Example 2

[0330] 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."

[0331] It is difficult for elderly and foreign patients to communicate smoothly with medical professionals, especially when it comes to accurately recognizing emotions and providing appropriate responses. Furthermore, in multilingual medical communication, it is necessary to ensure the accuracy and reliability of translations. Furthermore, the security of medical information is also a major issue.

[0332] 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.

[0333] In this invention, the server includes a means for transmitting voice data in real time, a means for performing noise reduction and normalization on the preprocessed voice data, and a means for analyzing the preprocessed voice data to recognize emotions. This enables medical communication that takes into account the user's emotions while overcoming language barriers. It also improves the accuracy and reliability of translations and ensures the security of medical information.

[0334] "Means for collecting voice data" refers to the function of capturing voices spoken by patients or medical professionals using the device's microphone.

[0335] The "means for transmitting audio data to a server in real time" refers to a mechanism for converting captured audio data into a digital format in real time and transmitting it to a server via a network.

[0336] The "noise reduction and normalization means" is a function that removes background noise from the received audio data and performs processing to equalize the volume.

[0337] "Means for recognizing emotions" refers to a technology that analyzes pre-processed voice data and identifies the user's emotional state from the intonation and tempo of the voice.

[0338] A "speech recognition model" is a technology that analyzes voice data and converts it into text format, often using automatic speech recognition (ASR) algorithms.

[0339] A "multilingual generative AI model" is an AI technology that supports multiple languages ​​and translates input text data into other languages.

[0340] "Display and playback means" refers to the functionality of providing the translated text and emotional information to the user visually and audibly.

[0341] "Means for expert checking and correction" refers to a system in which experts can check and correct the content of translated text and the results of emotion recognition.

[0342] "Means for encrypting data" refers to a method of protecting all data with advanced encryption technology, ensuring that data is sent and received securely during communication.

[0343] "Training measures to take cultural differences into account" refers to the learning process that generative AI models undergo to understand the cultural nuances of each language and translate or respond appropriately.

[0344] This invention is a system that enables smooth communication between elderly and foreign patients and medical professionals, and also enables emotion recognition. This system provides voice, text, translation, and real-time medical information to support accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it becomes possible to diagnose and treat patients taking into account their emotional state.

[0345] Basic configuration

[0346] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, a terminal that displays and plays back the converted and translated information, and an interface for experts to check and complete the translation results and recognized emotions. The system also employs data encryption and confidentiality technologies to meet high security standards.

[0347] Hardware and Software

[0348] 1. Device:

[0349] It is a device with a user interface and a built-in microphone for collecting and capturing voice data. The device also has a processor, memory, and a network module, and transmits the voice data to a server in real time.

[0350] 2. Server:

[0351] The server preprocesses the received voice data and performs emotion and speech recognition. It has a high-performance processor, memory, and data storage, and runs speech recognition models and generative AI models, which then perform noise reduction and data analysis.

[0352] 3. Software:

[0353] The speech recognition model running on the server uses an automatic speech recognition (ASR) algorithm to convert preprocessed speech data into text. The generative AI model is multilingual and translates text data into other languages. In particular, it employs generative AI technology that has a proven track record as a translation engine.

[0354] 4. Emotion Engine:

[0355] This engine identifies the user's emotions by analyzing the intonation and waveform of voice data, and recognizes a variety of emotional states (e.g., anger, sadness, anxiety).

[0356] 5. Security Technology:

[0357] The AES encryption algorithm is used to encrypt data, and data is sent and received securely via the TLS protocol during communication.

[0358] 6. Expert interface:

[0359] This interface allows experts to review the translation results and the perceived sentiment, and provide corrections or additional information as needed. Users can input appropriate actions, thereby improving the accuracy and reliability of the translation.

[0360] Specific examples

[0361] For a specific scenario, use the following prompt:

[0362] Prompt Sentence Examples

[0363] The patient is expected to say to the device in Japanese, "I have chest pain," and the doctor will understand this in English.

[0364] The device captures the patient's voice saying "I feel chest pain," converts it into digital format, and sends it to the server. The server performs noise reduction on the received voice data and then uses an emotion engine to recognize the emotion "anxiety." Next, a speech recognition model converts the voice into text, generating the text data "I feel chest pain." This text data is translated into English by a generative AI model, resulting in the result "I feel chest pain." The translation result and emotion information are displayed on the device and, if necessary, played back using a speech synthesis function. Finally, an expert reviews the result and makes corrections or asks additional questions as necessary. The expert can then ask an additional question, "When did the chest pain start?" and provide measures to alleviate the patient's anxiety.

[0365] As described above, this system not only enables fast and accurate medical communication across language barriers, but also provides a humane response that takes into account the patient's emotional state.

[0366] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0367] Step 1:

[0368] Audio data collection

[0369] Device: The patient or medical professional speaks into the device's microphone. Voice data is input in analog format, and the microphone captures the voice and converts it into digital format. Specifically, the device activates the microphone and records the patient saying, "I have a headache." The recorded voice data is stored in a buffer.

[0370] Input: Patient speech (e.g., "I have a headache")

[0371] Output: Digital audio data

[0372] Step 2:

[0373] Sending audio data

[0374] Terminal: Converts recorded audio data into digital format in real time and streams it to the server over the network. Specifically, the terminal packs audio data into packets and sends them to the server via the network module.

[0375] Input: Recorded digital audio data

[0376] Output: Audio data sent to the server

[0377] Step 3:

[0378] Audio data preprocessing

[0379] Server: Performs noise reduction and normalization on the received audio data to convert it into a clear, uniform format. Specifically, the server filters the audio data, removes background noise, and normalizes volume fluctuations.

[0380] Input: Audio data sent to the server

[0381] Output: Noise-reduced and normalized audio data

[0382] Step 4:

[0383] Emotion recognition by emotion engine

[0384] Server: Analyzes the preprocessed voice data and identifies the user's emotions from the intonation and waveform of the voice data. Specifically, the emotion engine analyzes the frequency components of the voice data and recognizes emotions such as "anger," "sadness," and "surprise."

[0385] Input: Preprocessed audio data

[0386] Output: Recognized emotion tag (e.g., "anxiety")

[0387] Step 5:

[0388] Speech-to-text conversion

[0389] Server: Using a speech recognition model, the preprocessed speech data is converted into text format. Specifically, the speech data is broken down into phonemes, which are then reconstructed into strings of characters using a language model to generate text data.

[0390] Input: Preprocessed audio data

[0391] Output: Text data (e.g., "I have a headache")

[0392] Step 6:

[0393] Text translation

[0394] Server: Translates text data into other languages ​​using a generative AI model. Specifically, the generative AI model first analyzes the input text and then generates a translation based on the grammar and vocabulary of the target language. For example, "I have a headache" is translated into "I have a headache."

[0395] Input: Text data (e.g., "I have a headache")

[0396] Output: Translated text data (e.g., "I have a headache")

[0397] Step 7:

[0398] Display and playback of translation results and emotions

[0399] Terminal: The translated text and emotion information are displayed on the user's terminal and played back using a speech synthesis function as needed. Specifically, the terminal displays the text on the screen and uses a speech synthesis engine to play back the translation result aloud. For example, the translation result is displayed on the doctor's terminal and the phrase "I have a headache" is played back.

[0400] Input: Translated text data and recognized sentiment tags

[0401] Output: Text displayed on the screen, translation results played back aloud, and emotional information displayed

[0402] Step 8:

[0403] Checked and supplemented by experts

[0404] User (medical professional): Checks the translated content and the recognized sentiment, and provides corrections or additional information as needed. Specifically, the professional reviews the displayed translation, makes corrections if necessary, and enters a follow-up question. For example, a doctor might enter a follow-up question such as, "What causes headaches?"

[0405] Input: Translated text data and recognized sentiment tags

[0406] Output: Corrected translation and additional questions

[0407] Step 9:

[0408] Data encryption and protection

[0409] Server: All data is protected with advanced encryption technology. Specifically, voice and text data is encrypted with the AES encryption algorithm, and data in transit is securely transmitted and received using the TLS protocol.

[0410] Input: Audio data, text data, emotion data

[0411] Output: AES encrypted data, secure communication via TLS protocol

[0412] Step 10:

[0413] Training a generative AI model

[0414] Server: Trains the AI ​​model to understand cultural differences and multilingual environments. Specifically, the AI ​​model learns using multilingual datasets to improve its ability to understand Japanese nuances and cultural background.

[0415] Input: Multilingual dataset

[0416] Output: Generative AI models that are culturally and multilingually responsive

[0417] (Application example 2)

[0418] 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."

[0419] Elderly and foreign customers can sometimes find it difficult to communicate smoothly with store staff. In such situations, customers may not receive appropriate service due to language barriers or a lack of understanding of their emotions. Particularly in physical stores, where appropriate responses are required in real time, a system that solves these issues is necessary.

[0420] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generation AI model, means for displaying the translated text and emotional state, and means for audibly playing back the translated text and emotional state. This makes it possible to translate voice data into multiple languages ​​and display and play back the emotional state in real time, allowing elderly people and foreign customers to receive appropriate services.

[0421] "Voice data" is a digital recording of a user's speech.

[0422] "Collection means" refers to the devices and techniques used to collect audio data.

[0423] "Preprocessing" refers to performing data processing such as noise removal and normalization on collected voice data.

[0424] A "speech recognition model" is an algorithm or machine learning model that analyzes voice data and converts it into text data.

[0425] "Text" is character data converted from audio data.

[0426] A "multilingual generative AI model" is an artificial intelligence model used to translate text input in one language into another language.

[0427] "Translation" is the conversion of text written in one language into another.

[0428] "Display means" refers to devices or techniques for visually presenting the translated text and emotional state to the user.

[0429] "Audio playback" refers to synthesizing text data into audio and providing it to the user audibly.

[0430] The "emotional state" indicates the emotional state recognized from the user's speech and facial expressions.

[0431] "Means of expert checking and supplementation" refers to a method in which an expert checks the accuracy of the translation results and provides corrections or additional information as necessary.

[0432] "Encryption" is a technology that converts data based on a specific algorithm to prevent unauthorized access by third parties.

[0433] "Cultural training" is the process by which generative AI models learn to adapt to multicultural environments.

[0434] This invention is a system for facilitating communication between elderly and foreign customers and store staff in brick-and-mortar stores. This system collects user speech data and performs preprocessing, translation, and emotion recognition using the following means:

[0435] 1. Audio data collection:

[0436] The server uses a terminal to collect voice data, such as a microphone built into smart glasses or a smartphone, which allows the server to capture customer speech in real time.

[0437] 2. Preprocessing the audio data:

[0438] The server then denoises and normalizes the received audio data using voice processing APIs from Amazon AWS and Google Cloud, making the data easier to analyze.

[0439] 3. Audio to text conversion:

[0440] The server then feeds the preprocessed audio data into a speech recognition model and converts it into text using an automatic speech recognition algorithm such as IBM Watson Speech to Text.

[0441] 4. Text translation:

[0442] The server translates the converted text using a multilingual generative AI model, using a translation engine such as the DeepL translation API, enabling smooth communication between different languages.

[0443] 5. Emotion recognition:

[0444] The server uses a natural language processing API (NLP) from Google Cloud to analyze the user's emotions from the text data. By identifying the user's emotional state, it can provide a more appropriate response.

[0445] 6. Displaying translation results and emotional state and playing audio:

[0446] The translated text and emotional state are displayed on the HUD (heads-up display) or smartphone screen, and the translation is played back via speech synthesis for real-time understanding by both the customer and store staff.

[0447] 7. Expert check and completion:

[0448] Experts check the accuracy of the translated content and provide corrections or additional information as needed, which increases the reliability of the system.

[0449] For example, when a store clerk asks, "Welcome. What can I do for you?", an elderly customer responds, "My feet have been hurting recently, so it's a bit difficult to walk." This speech data is collected, preprocessed, and converted into text using a speech recognition model. Next, this text is translated into English using the DeepL translation API, and the emotional state is identified as "negative" using Google Cloud's Natural Language API. Finally, the HUD displays, "My feet have been hurting recently, so it's a bit difficult to walk. Emotion score: -0.3," and the translation result is conveyed to the store clerk through audio playback.

[0450] An example of a prompt sentence for the generative AI model is as follows:

[0451] "Audio of a patient speaking Japanese in a hospital"

[0452] "output_language: English"

[0453] "output_emotion: emotion analysis"

[0454] "expected_output: Translation result: 'My feet have been hurting recently, so it's a bit difficult to walk.' Sentiment score: '-0.3'"

[0455] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0456] Step 1:

[0457] The user speaks into the microphone built into the smart glasses or smartphone. The device captures the customer's voice data and transmits it to the server in real time. The input is the customer's voice data, and the output is the digital voice data sent to the server.

[0458] Step 2:

[0459] The server denoises and normalises the received audio data. This process uses the audio processing APIs of Amazon AWS and Google Cloud. The input is the received audio data, and the output is the denoised and normalised audio data.

[0460] Step 3:

[0461] The server inputs the preprocessed audio data into a speech recognition model and converts it into text. IBM Watson Speech to Text is used here. The input is the preprocessed audio data, and the output is text data.

[0462] Step 4:

[0463] The server translates the converted text using a multilingual generative AI model. This process uses the DeepL translation API. The input is text data, and the output is the translated text.

[0464] Step 5:

[0465] The server uses a natural language processing API to analyze the user's emotional state from the translated text. Google Cloud's Natural Language API is used here. The input is the translated text data, and the output is the emotion score and emotion type.

[0466] Step 6:

[0467] The server displays the translated text and emotional state on a HUD (Heads-Up Display) or smartphone screen. The displayed content includes the translated text and the emotional score. The input is the translated text and the emotional state, and the output is the displayed information.

[0468] Step 7:

[0469] The server synthesizes the translation result and plays it back, allowing the user (store clerk) to listen to the translation in real time. The input is the translated text, and the output is audio data.

[0470] Step 8:

[0471] The expert (clerk) checks the displayed and played translation results and emotional state, and provides corrections or additional information as needed. The input is the translated text and emotional state, and the output is the corrected or supplemented information.

[0472] 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.

[0473] 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.

[0474] 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.

[0475] [Second embodiment]

[0476] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0477] 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.

[0478] 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).

[0479] 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.

[0480] 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.

[0481] 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).

[0482] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0483] 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.

[0484] 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.

[0485] 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.

[0486] 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.

[0487] 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."

[0488] This is a system to support smooth communication between elderly and foreign patients and medical professionals. The system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment.

[0489] Program implementation and processing explanation

[0490] Basic configuration

[0491] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[0492] Overview of the main measures that will emerge

[0493] 1. Collection of audio data

[0494] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[0495] 2. Preprocessing of audio data

[0496] Server: Performs noise reduction and normalization on the received audio data, removing unnecessary noise and converting the audio data into a format that is easier to analyze.

[0497] 3. Speech to text conversion

[0498] Server: The preprocessed speech data is analyzed using a speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[0499] 4. Text Translation

[0500] Server: The converted text data is translated in real time using a multilingual generative AI model. For example, the Japanese text "I have a headache" is translated into English as "I have a headache."

[0501] 5. Display and playback of translated text and audio

[0502] On your device: The translated text will be displayed on your device screen and played back using speech synthesis, if available. For example, the translation will be displayed on your device and the speech synthesis will say "I have a headache."

[0503] 6. Expert check and completion

[0504] User (medical professional): Checks the translation results and corrects or completes them as necessary. This improves the accuracy and reliability of the translation. For example, a medical professional corrects "windpipe" to "trachea."

[0505] 7. Data Encryption and Protection

[0506] Server: All data is protected with advanced encryption technology, for example, using the AES encryption algorithm, and data in transit is sent and received over secure protocols (TLS / SSL).

[0507] 8. Training a generative AI model

[0508] Server: Generative AI models are trained to understand cultural differences and sensitivities, for example, using datasets specific to Japanese customs and expressions.

[0509] Specific examples

[0510] Consider a scenario in which a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital.

[0511] 1. Voice input: The patient speaks to the terminal in English, "I have a headache."

[0512] 2. Data preprocessing and speech recognition: This speech data is sent to the server, where it is noise-reduced and converted to text using an ASR model.

[0513] 3. Text translation: A generative AI model translates the English phrase "I have a headache" into Japanese phrase "I have a headache."

[0514] 4. Display and audio output: The translated Japanese text is displayed on the doctor's device and played as audio if necessary.

[0515] 5. Expert review: A doctor reviews the translation and asks additional questions or makes a diagnosis.

[0516] The above steps will enable rapid and accurate medical communication that transcends language barriers.

[0517] The processing flow will be explained below.

[0518] Step 1:

[0519] Audio data collection

[0520] Device: The patient or healthcare professional speaks into the device. The device's built-in microphone captures the audio and begins recording.

[0521] Example of operation: When a patient says "I have a headache" to the device, the voice is recorded.

[0522] Step 2:

[0523] Sending audio data

[0524] Terminal: Streams recorded audio data to the server in real time.

[0525] How it works: Captured audio data is converted into digital form and sent over a network to a server.

[0526] Step 3:

[0527] Audio data preprocessing

[0528] Server: Performs noise reduction and normalization on the received audio data, converting it into a clear and consistent format.

[0529] How it works: A filtering algorithm is applied to remove background noise and even out volume variations.

[0530] Step 4:

[0531] Text conversion using speech recognition models

[0532] Server: The preprocessed speech data is input into the speech recognition model and converted into text.

[0533] Example of how it works: A speech recognition algorithm analyzes the characteristics of a voice and converts the speech "I have a headache" into the text "I have a headache."

[0534] Step 5:

[0535] Text translation

[0536] Server: Using a generative AI model, translate text data into the target language. Use a multilingual translation engine.

[0537] Example of how it works: The text "I have a headache" is translated into English as "I have a headache".

[0538] Step 6:

[0539] Display and playback of translation results

[0540] On Device: Displays translated text and audio on the user's device, optionally using speech synthesis to play the translation aloud.

[0541] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache" in English.

[0542] Step 7:

[0543] Checked and supplemented by experts

[0544] User (medical professional): Review the translated content and provide corrections or additional information as needed.

[0545] Example of how it works: A doctor checks the translation results and corrects any inaccuracies. For example, he or she asks an additional question in Japanese: "Please tell me the cause of my headache."

[0546] Step 8:

[0547] Data encryption and protection

[0548] Server: We encrypt all data sent and received and use advanced security protocols to prevent unauthorized access to your data by third parties.

[0549] How it works: The server encrypts data using the AES encryption algorithm, and the encrypted data is sent and received via the TLS protocol.

[0550] Step 9:

[0551] Training a generative AI model

[0552] Server: Trains AI models to understand cultural differences and multilingual environments, thereby reducing mistranslations and cultural misunderstandings.

[0553] How it works: An AI model trains on a server using a multilingual dataset to improve its ability to understand Japanese-specific nuances.

[0554] Example 1

[0555] 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."

[0556] In today's medical settings, communication barriers between elderly or foreign patients and medical professionals are a serious problem. It is particularly difficult to communicate smoothly with patients who have language differences, hearing impairments, or cultural backgrounds, which can delay appropriate diagnoses and treatment. A system that can resolve these issues and support accurate and prompt communication in medical settings is needed.

[0557] 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.

[0558] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for experts to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for noise reduction and normalization, means for inputting voice into a voice recognition model, and means for synthesizing voice, thereby enabling accurate and prompt communication between patients and medical professionals with different languages, hearing impairments, or cultural backgrounds.

[0559] An "audio data collection means" is a device or method for capturing and transmitting audio from a patient or medical professional to a server.

[0560] The "audio data preprocessing means" refers to a device or method for converting collected audio data into a form that is easy to analyze by performing processes such as noise reduction and normalization.

[0561] A "speech recognition model" is a machine learning algorithm that analyzes preprocessed speech data and converts it into text data.

[0562] A "multilingual generative AI model" is an artificial intelligence model used to translate between different languages ​​in real time.

[0563] "Display and playback means" refers to a device or method for displaying the translated text on the screen of the terminal and, if necessary, synthesizing it to speech and playing it back.

[0564] "Expert checking and completion means" refers to an interface and method for medical professionals to check the translation results and make corrections or completions as necessary.

[0565] A "data encryption means" is a device or method for encrypting data using an encryption algorithm to protect all data processed within the system.

[0566] A "means for training a generative AI model" is a device or method for training a generative AI model with a particular dataset so that the model understands cultural differences and sensitivities.

[0567] "Noise reduction and normalization means" refers to a device or method for reducing noise in audio data during pre-processing and adjusting the volume and pitch to a consistent level.

[0568] A "means for inputting speech into a speech recognition model" is a device or method for appropriately inputting preprocessed speech data into a speech recognition model.

[0569] "Speech synthesis means" refers to a device or method for generating speech based on text data and reproducing it.

[0570] This invention relates to a system that supports smooth communication between elderly and foreign patients and medical professionals. Specifically, it enables accurate and prompt medical treatment by providing voice, text, translation, and real-time medical information.

[0571] Fundamental System Configuration

[0572] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[0573] Technical Details

[0574] Hardware and Basic Operation

[0575] The device is used by the patient or a medical professional and collects voice data through a built-in microphone. The device can be a smartphone, tablet, or dedicated voice collection device. The collected voice data is sent to a server in real time. For example, if a patient says, "I have a headache," the voice is captured instantly.

[0576] The server preprocesses the received audio data. The main processes are noise reduction and normalization. Noise reduction removes background noise, and normalization maintains the quality of the audio data.

[0577] The preprocessed speech data is then converted into text data using an Automatic Speech Recognition (ASR) model. For example, the Japanese phrase "I have a headache" is converted directly into text.

[0578] The converted text is then translated in real time by a multilingual generative AI model, pre-trained to understand cultural differences and sensitivities, for example, translating the Japanese text "Watashi wa tou ga kaiyuu" (I have a headache) into English.

[0579] The translated text and audio are then displayed and played back on the device. The translated text is displayed on the device screen and the translation is played back using speech synthesis. For example, the translated "I have a headache" is displayed and played back as audio.

[0580] The user (medical professional) can check the translation results through the interface provided by the system and make corrections or additions as necessary, thereby improving the accuracy and reliability of the translation. For example, a medical professional can correct "windpipe" to "trachea."

[0581] The server protects all data with advanced encryption technology, specifically using the AES encryption algorithm, and data in transit is sent and received via secure protocols (TLS / SSL).

[0582] Specific examples

[0583] Consider a case where a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital. The following steps are taken:

[0584] 1. Voice input:

[0585] The patient speaks to the terminal in English, "I have a headache."

[0586] 2. Data preprocessing and speech recognition:

[0587] This audio data is sent to a server, where it is noise reduced and normalized before being converted to text using an ASR model.

[0588] 3. Text Translation:

[0589] The server uses a generative AI model to translate the English phrase "I have a headache" into Japanese phrase "I have a headache."

[0590] 4. Display and audio output:

[0591] The translated Japanese text is displayed on the doctor's device and the audio is played back as "I have a headache."

[0592] 5. Expert review:

[0593] The doctor will check the translated content and provide appropriate diagnosis and treatment.

[0594] Prompt Sentence Examples

[0595] Examples of prompts to enter into this system are:

[0596] "If a patient says, 'I have a headache,' in English, convert that audio data into Japanese text and speech for the doctor to review."

[0597] The above procedures and system configuration will enable rapid and accurate medical communication that transcends language barriers.

[0598] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0599] Step 1:

[0600] Audio data collection

[0601] The device captures the voice of the patient or medical professional. When the patient speaks into the device, the built-in microphone collects the voice and transmits it to the server in real time. The input is voice data, and the output is raw voice data sent to the server.

[0602] Specific behavior:

[0603] The patient speaks into the terminal's microphone, "I have a headache."

[0604] The device collects the voice data and sends it to the server.

[0605] Step 2:

[0606] Audio data preprocessing

[0607] The server performs noise reduction and normalization on the received audio data, converting it into a form that is easier to analyze. The input is raw audio data, and the output is preprocessed audio data.

[0608] Specific behavior:

[0609] The server receives the audio data.

[0610] Apply a noise reduction algorithm to remove background noise.

[0611] Normalizes the volume and pitch of audio data.

[0612] Step 3:

[0613] Speech-to-text conversion

[0614] The server inputs the preprocessed speech data into an Automatic Speech Recognition (ASR) model and converts it into text data. The input is the preprocessed speech data, and the output is text data.

[0615] Specific behavior:

[0616] The server inputs the speech data into the ASR model.

[0617] The ASR model analyzes the speech data and generates the text "I have a headache."

[0618] Step 4:

[0619] Text translation

[0620] The server translates the converted text using a generative AI model. The input is text data, and the output is translated text data.

[0621] Specific behavior:

[0622] The server inputs text into the generative AI model.

[0623] A generative AI model translates the Japanese text "I have a headache" into English "I have a headache."

[0624] Step 5:

[0625] View and play translated text and audio

[0626] The terminal displays the translated text received from the server on the screen and plays it back using speech synthesis if necessary. The input is the translated text data, and the output is the displayed text and speech data.

[0627] Specific behavior:

[0628] The terminal receives the translation result from the server.

[0629] The translated text is displayed on the screen.

[0630] Uses a speech synthesis engine to play the translated text aloud.

[0631] Step 6:

[0632] Checked and supplemented by experts

[0633] The user (medical professional) checks the translation results displayed on the terminal and makes corrections or completions as necessary. The input is the translated text data, and the output is the corrected or completed text data.

[0634] Specific behavior:

[0635] The medical professional checks the translation results displayed on the device's display.

[0636] If the translation is not accurate, correct it.

[0637] Step 7:

[0638] Data encryption and protection

[0639] The server encrypts all audio and text data. The input is the processed data and the output is the encrypted data.

[0640] Specific behavior:

[0641] The server encrypts the audio and text data using the AES encryption algorithm.

[0642] Encrypted data is sent and received using secure protocols (TLS / SSL).

[0643] Step 8:

[0644] Training a generative AI model

[0645] The server trains the generative AI model to understand cultural differences and sensitivities: the input is a specific culture and language dataset, and the output is an optimized generative AI model.

[0646] Specific behavior:

[0647] The server inputs specific culture and language datasets into the AI ​​model.

[0648] The server optimizes the parameters of the AI ​​model and performs training.

[0649] (Application example 1)

[0650] 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."

[0651] To ensure smooth communication between elderly and foreign patients and medical professionals, accurate real-time speech recognition and translation, as well as secure processing and transmission of the data, are necessary. However, current systems lack the ability to process voice data in real time, securely encrypt the data, and train generative AI models that take cultural differences into account, hindering accurate and prompt medical responses.

[0652] 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.

[0653] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a speech recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for an expert to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for securely storing and transmitting the encrypted data, and means for processing the voice data collected by the terminal in real time, thereby enabling real-time and accurate medical communication between patients and medical professionals.

[0654] "Audio data" refers to data in which an audio signal is recorded as digital information.

[0655] A "collection means" is a device or system for acquiring and transmitting audio data to another processing unit.

[0656] The "preprocessing means" is a device or system that performs preprocessing such as noise removal and normalization on collected voice data.

[0657] A "voice recognition model" is an algorithm or AI model that analyzes voice data and converts it into a string of characters (text).

[0658] "Text" is digital data expressed as character information.

[0659] A "multilingual generative AI model" is an AI model that supports multiple languages ​​and translates from one language to another in real time.

[0660] A "translation tool" is a device or system that converts text in a particular language into another language.

[0661] "Display means" refers to a device or interface for visually showing the translation results to the user.

[0662] "Playback means" refers to a device or system that provides the translation result to the user as audio.

[0663] "Means for expert checking and correction" refers to a device or system that allows an expert to verify the accuracy of the translation results and correct them if necessary.

[0664] An "encryption means" is a device or system that encrypts data to protect the security of the data.

[0665] A "training means for a generative AI model to take cultural differences into account" is a device or system that trains a generative AI model to understand differences in cultural background and expression.

[0666] A "real-time processing means" is a device or system that instantly analyzes and processes audio data and provides results to the user in real time.

[0667] "Storage and transmission means" means a device or system that securely stores encrypted data and transmits it wherever needed.

[0668] To implement this invention, a system must be constructed by combining a series of hardware and software. The specific system configuration, processing procedures, and usage methods are described below.

[0669] System configuration

[0670] This system is a multi-function system for speech recognition, real-time translation, data encryption, and displaying and playing back the results.

[0671] Hardware used

[0672] 1. Collection device: A device for collecting voice data, such as a smartphone or head-mounted display. This device has a built-in highly sensitive microphone.

[0673] 2. Servers: High-performance servers are used for data pre-processing, speech recognition, text conversion, translation, and data encryption, including cloud servers.

[0674] 3. Display device: A device that displays the translated data and plays it back as audio. This includes smartphones and tablets.

[0675] Software used

[0676] 1. Speech recognition library: A library for converting voice data to text (e.g., speech_recognition).

[0677] 2. Translation library: A library for translating text into another language (e.g. googletrans).

[0678] 3. Cryptography library: A library for performing data encryption (e.g., the Fernet module for cryptography).

[0679] System processing explanation in natural language

[0680] The server receives voice data from a collection terminal, which is used by a patient or a medical professional to capture the voice data. The captured voice data undergoes pre-processing for noise reduction and normalization before being sent to the server.

[0681] The server converts the preprocessed speech data into text using a speech recognition library. This text data is translated into the specified language in real time using a multilingual generative AI model. For example, if a patient says "I have a headache" in English, this will be translated into Japanese as "I have a headache."

[0682] The translated text is sent to a display terminal where it is displayed on the screen and played back as audio. The system also includes an interface that allows experts to check the accuracy of the translation and make corrections or additions if necessary.

[0683] All data is encrypted using advanced encryption technology and transmitted over secure protocols. The server uses the AES encryption algorithm and TLS / SSL protocols to ensure data security.

[0684] Examples of concrete examples and prompts

[0685] Specific usage scenarios include the following:

[0686] 1. Example: When an English-speaking patient visits a Japanese hospital, they use their smartphone to communicate their symptoms to the doctor. When they speak into their smartphone, for example, "I have a headache," this voice data is translated into Japanese and displayed and played on the doctor's device.

[0687] 2. Example prompt:

[0688] I'm an English-speaking patient with a headache. I say "I have a headache" in English to my smartphone. Please translate the voice data into Japanese and display it so that my doctor can see it. Also, please encrypt the data to ensure its security.

[0689] This invention enables real-time and accurate medical communication between elderly and foreign patients and medical professionals.

[0690] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0691] Step 1:

[0692] The collection terminal captures the speech data uttered by the patient or medical professional. The input is the speech signal collected from the microphone of the collection terminal. This speech signal is converted into a digital format and subjected to noise reduction and normalization processing. The output is the pre-processed speech data.

[0693] Step 2:

[0694] The preprocessed voice data is sent from the collection terminal to the server. The input is the preprocessed voice data. The collection terminal sends the data to the server using a transmission protocol. The output is the voice data received at the server side.

[0695] Step 3:

[0696] The server performs speech recognition processing on the received audio data. The input is noise-reduced and normalized audio data. The server uses a speech recognition library (e.g., speech_recognition) to convert the audio data into text data. The output is text data.

[0697] Step 4:

[0698] The server translates the converted text data using a multilingual generative AI model. The input is text data generated by a speech recognition model. A translation library (e.g., GoogleTrans) is used to translate the text into the specified language in real time. The output is the translated text data.

[0699] Step 5:

[0700] The server encrypts the translated text data. The input is the translated text data. The data is encrypted using an encryption library (e.g. Fernet). The output is the encrypted text data.

[0701] Step 6:

[0702] Encrypted text data is sent from the server to the display terminal. The input is the encrypted text data. The server transfers the data using a secure protocol (e.g., TLS / SSL). The output is the encrypted text data received by the display terminal.

[0703] Step 7:

[0704] The display terminal decrypts encrypted text data received from the server. The input is the encrypted text data. The display terminal uses the encryption library to decrypt the data. The output is the decrypted text data.

[0705] Step 8:

[0706] The display terminal displays and plays back the decoded text data. The input is the decoded text data. The user interface displays the text data on the screen and uses speech synthesis to play the text aloud. The output is a translation result that can be recognized visually and audibly by the user.

[0707] Step 9:

[0708] The expert user uses a display terminal to check the accuracy of the translated text data and make corrections or completions as necessary. The input is the decoded and displayed text data. The user makes corrections through the interface, and the results are reflected in the system. The output is the corrected and completed accurate translation data.

[0709] This series of steps enables accurate and secure medical communication in real time.

[0710] 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.

[0711] This invention is a system for supporting smooth communication and emotion recognition between elderly and foreign patients and medical professionals. This system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it is possible to make diagnoses and treatments that take into account the patient's emotional state.

[0712] Program implementation and processing explanation

[0713] Basic configuration

[0714] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, and a terminal that displays and plays back the converted and translated information. An interface is also provided for experts to check and complete the translation results and recognized emotions. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[0715] Overview of the main measures that will emerge

[0716] 1. Collection of audio data

[0717] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[0718] 2. Sending audio data

[0719] Terminal: Streams recorded audio data to the server in real time. The captured audio data is converted into a digital format and sent over the network to the server.

[0720] 3. Preprocessing of audio data

[0721] Server: Noise reduction and normalization processes are performed on the received audio data, converting the audio into a clear and consistent format. Filtering algorithms are applied to remove background noise and even out volume variations.

[0722] 4. Emotion Recognition by Emotion Engine

[0723] Server: Analyzes the preprocessed audio and video data to recognize the user's emotional state. For example, it identifies emotions such as "anger," "sadness," and "surprise." The emotion recognition algorithm reads the user's emotions from their voice intonation and facial expressions.

[0724] 5. Speech to Text Conversion

[0725] Server: The preprocessed speech data is input into the speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[0726] 6. Text Translation

[0727] Server: Translates text data into the target language using a generative AI model. It uses a multilingual translation engine. For example, the text "I have a headache" is translated into English as "I have a headache."

[0728] 7. Display and playback of translation results and emotions

[0729] Terminal: The translated text and recognized emotion information are displayed on the user's terminal. If necessary, the translation result is played back aloud using speech synthesis. For example, the translation result is displayed on the doctor's terminal, and the speech synthesis reproduces "I have a headache" in English. The display screen also indicates the patient's emotional state (e.g., "anxiety").

[0730] 8. Expert check and completion

[0731] User (medical professional): Review the translated content and the perceived sentiment, and provide corrections or additional information as needed. This improves the accuracy and reliability of the translation. For example, a doctor may review the translation and ask a follow-up question such as "What is causing your headache?", as well as provide suggestions to ease the patient's anxiety.

[0732] 9. Data Encryption and Protection

[0733] Server: All data is protected with advanced encryption technology. Voice and text data is encrypted with the AES encryption algorithm, and data in transit is also securely transmitted via the TLS protocol.

[0734] 10. Training a generative AI model

[0735] Server: Trains AI models to understand cultural differences and multilingual environments. Using multilingual datasets, AI models learn and improve their ability to understand Japanese-specific nuances.

[0736] These steps not only enable fast and accurate medical communication across language barriers, but also provide a more humane response that takes into account the patient's emotional state.

[0737] The processing flow will be explained below.

[0738] Step 1:

[0739] Audio data collection

[0740] Device: The patient or healthcare professional speaks into the device, which has a built-in microphone that captures the voice and records the data in real time.

[0741] Example of operation: When a patient says "I have a headache" to the device, the voice is instantly recorded and saved as digital audio data.

[0742] Step 2:

[0743] Sending audio data

[0744] Terminal: Recorded audio data is sent to the server in streaming format.

[0745] How it works: Send captured audio data to a server using any Wi-Fi or 4G network within range.

[0746] Step 3:

[0747] Audio data preprocessing

[0748] Server: Performs noise reduction and normalization on the received audio data, making it easier to analyze.

[0749] How it works: Using a noise filtering algorithm on the server side to remove background noise and enhance the audio signal.

[0750] Step 4:

[0751] Emotion recognition by emotion engine

[0752] Server: Analyzes the pre-processed audio and video data (if necessary) and recognizes the user's emotions. An emotion recognition algorithm is used.

[0753] Example of how it works: The server analyzes the intonation, speed, and emphasized parts of the voice to identify emotions such as "anger," "sadness," and "surprise." It also uses facial expression data recognized using a camera.

[0754] Step 5:

[0755] Speech-to-text conversion

[0756] Server: The preprocessed speech data is input into the speech recognition model (ASR) and converted into corresponding text data.

[0757] Example of how it works: A speech recognition algorithm listens for speech characteristics and converts the speech "I have a headache" into text "I have a headache."

[0758] Step 6:

[0759] Text translation

[0760] Server: Using a generative AI model, the converted text data is translated into the required language using a multilingual translation engine.

[0761] Working example: Translate the Japanese phrase "I have a headache" into English.

[0762] Step 7:

[0763] Display and playback of translation results and emotional information

[0764] Device: The translated text and recognized emotion information are displayed on the user's device, and the translation result is played aloud using speech synthesis if necessary.

[0765] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache." Emotional information such as "Emotion: Anxiety" is also displayed on the display screen.

[0766] Step 8:

[0767] Checked and supplemented by experts

[0768] User (medical professional): Checks the content of the translated text and sentiment information, and makes corrections or asks additional questions as needed, thereby improving the accuracy of the diagnosis.

[0769] Example of how it works: A doctor checks the translation results and emotional information, and then asks an additional question in Japanese, such as "Please tell me the cause of my headache." The system translates this into English in real time.

[0770] Step 9:

[0771] Data encryption

[0772] Server: All data sent and received is protected with the AES encryption algorithm, and communication is carried out via the TLS / SSL protocol.

[0773] Example of operation: All data, including recorded voice data, translated text, and emotional information, is encrypted and transmitted over a secure communication channel.

[0774] Step 10:

[0775] Training a generative AI model

[0776] Server: Trains generative AI models to understand cultural differences and multilingual environments, using multilingual datasets for training.

[0777] How it works: The server trains an AI model using a dataset on Japanese culture to understand Japanese-specific expressions and nuances.

[0778] The above processing flow enables accurate and prompt communication between patients and medical professionals, and also enables responses that take into account the patient's emotional state.

[0779] Example 2

[0780] 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."

[0781] It is difficult for elderly and foreign patients to communicate smoothly with medical professionals, especially when it comes to accurately recognizing emotions and providing appropriate responses. Furthermore, in multilingual medical communication, it is necessary to ensure the accuracy and reliability of translations. Furthermore, the security of medical information is also a major issue.

[0782] 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.

[0783] In this invention, the server includes a means for transmitting voice data in real time, a means for performing noise reduction and normalization on the preprocessed voice data, and a means for analyzing the preprocessed voice data to recognize emotions. This enables medical communication that takes into account the user's emotions while overcoming language barriers. It also improves the accuracy and reliability of translations and ensures the security of medical information.

[0784] "Means for collecting voice data" refers to the function of capturing voices spoken by patients or medical professionals using the device's microphone.

[0785] The "means for transmitting audio data to a server in real time" refers to a mechanism for converting captured audio data into a digital format in real time and transmitting it to a server via a network.

[0786] The "noise reduction and normalization means" is a function that removes background noise from the received audio data and performs processing to equalize the volume.

[0787] "Means for recognizing emotions" refers to a technology that analyzes pre-processed voice data and identifies the user's emotional state from the intonation and tempo of the voice.

[0788] A "speech recognition model" is a technology that analyzes voice data and converts it into text format, often using automatic speech recognition (ASR) algorithms.

[0789] A "multilingual generative AI model" is an AI technology that supports multiple languages ​​and translates input text data into other languages.

[0790] "Display and playback means" refers to the functionality of providing the translated text and emotional information to the user visually and audibly.

[0791] "Means for expert checking and correction" refers to a system in which experts can check and correct the content of translated text and the results of emotion recognition.

[0792] "Means for encrypting data" refers to a method of protecting all data with advanced encryption technology, ensuring that data is sent and received securely during communication.

[0793] "Training measures to take cultural differences into account" refers to the learning process that generative AI models undergo to understand the cultural nuances of each language and translate or respond appropriately.

[0794] This invention is a system that enables smooth communication between elderly and foreign patients and medical professionals, and also enables emotion recognition. This system provides voice, text, translation, and real-time medical information to support accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it becomes possible to diagnose and treat patients taking into account their emotional state.

[0795] Basic configuration

[0796] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, a terminal that displays and plays back the converted and translated information, and an interface for experts to check and complete the translation results and recognized emotions. The system also employs data encryption and confidentiality technologies to meet high security standards.

[0797] Hardware and Software

[0798] 1. Device:

[0799] It is a device with a user interface and a built-in microphone for collecting and capturing voice data. The device also has a processor, memory, and a network module, and transmits the voice data to a server in real time.

[0800] 2. Server:

[0801] The server preprocesses the received voice data and performs emotion and speech recognition. It has a high-performance processor, memory, and data storage, and runs speech recognition models and generative AI models, which then perform noise reduction and data analysis.

[0802] 3. Software:

[0803] The speech recognition model running on the server uses an automatic speech recognition (ASR) algorithm to convert preprocessed speech data into text. The generative AI model is multilingual and translates text data into other languages. In particular, it employs generative AI technology that has a proven track record as a translation engine.

[0804] 4. Emotion Engine:

[0805] This engine identifies the user's emotions by analyzing the intonation and waveform of voice data, and recognizes a variety of emotional states (e.g., anger, sadness, anxiety).

[0806] 5. Security Technology:

[0807] The AES encryption algorithm is used to encrypt data, and data is sent and received securely via the TLS protocol during communication.

[0808] 6. Expert interface:

[0809] This interface allows experts to review the translation results and the perceived sentiment, and provide corrections or additional information as needed. Users can input appropriate actions, thereby improving the accuracy and reliability of the translation.

[0810] Specific examples

[0811] For a specific scenario, use the following prompt:

[0812] Prompt Sentence Examples

[0813] The patient is expected to say to the device in Japanese, "I have chest pain," and the doctor will understand this in English.

[0814] The device captures the patient's voice saying "I feel chest pain," converts it into digital format, and sends it to the server. The server performs noise reduction on the received voice data and then uses an emotion engine to recognize the emotion "anxiety." Next, a speech recognition model converts the voice into text, generating the text data "I feel chest pain." This text data is translated into English by a generative AI model, resulting in the result "I feel chest pain." The translation result and emotion information are displayed on the device and, if necessary, played back using a speech synthesis function. Finally, an expert reviews the result and makes corrections or asks additional questions as necessary. The expert can then ask an additional question, "When did the chest pain start?" and provide measures to alleviate the patient's anxiety.

[0815] As described above, this system not only enables fast and accurate medical communication across language barriers, but also provides a humane response that takes into account the patient's emotional state.

[0816] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0817] Step 1:

[0818] Audio data collection

[0819] Device: The patient or medical professional speaks into the device's microphone. Voice data is input in analog format, and the microphone captures the voice and converts it into digital format. Specifically, the device activates the microphone and records the patient saying, "I have a headache." The recorded voice data is stored in a buffer.

[0820] Input: Patient speech (e.g., "I have a headache")

[0821] Output: Digital audio data

[0822] Step 2:

[0823] Sending audio data

[0824] Terminal: Converts recorded audio data into digital format in real time and streams it to the server over the network. Specifically, the terminal packs audio data into packets and sends them to the server via the network module.

[0825] Input: Recorded digital audio data

[0826] Output: Audio data sent to the server

[0827] Step 3:

[0828] Audio data preprocessing

[0829] Server: Performs noise reduction and normalization on the received audio data to convert it into a clear, uniform format. Specifically, the server filters the audio data, removes background noise, and normalizes volume fluctuations.

[0830] Input: Audio data sent to the server

[0831] Output: Noise-reduced and normalized audio data

[0832] Step 4:

[0833] Emotion recognition by emotion engine

[0834] Server: Analyzes the preprocessed voice data and identifies the user's emotions from the intonation and waveform of the voice data. Specifically, the emotion engine analyzes the frequency components of the voice data and recognizes emotions such as "anger," "sadness," and "surprise."

[0835] Input: Preprocessed audio data

[0836] Output: Recognized emotion tag (e.g., "anxiety")

[0837] Step 5:

[0838] Speech-to-text conversion

[0839] Server: Using a speech recognition model, the preprocessed speech data is converted into text format. Specifically, the speech data is broken down into phonemes, which are then reconstructed into strings of characters using a language model to generate text data.

[0840] Input: Preprocessed audio data

[0841] Output: Text data (e.g., "I have a headache")

[0842] Step 6:

[0843] Text translation

[0844] Server: Translates text data into other languages ​​using a generative AI model. Specifically, the generative AI model first analyzes the input text and then generates a translation based on the grammar and vocabulary of the target language. For example, "I have a headache" is translated into "I have a headache."

[0845] Input: Text data (e.g., "I have a headache")

[0846] Output: Translated text data (e.g., "I have a headache")

[0847] Step 7:

[0848] Display and playback of translation results and emotions

[0849] Terminal: The translated text and emotion information are displayed on the user's terminal and played back using a speech synthesis function as needed. Specifically, the terminal displays the text on the screen and uses a speech synthesis engine to play back the translation result aloud. For example, the translation result is displayed on the doctor's terminal and the phrase "I have a headache" is played back.

[0850] Input: Translated text data and recognized sentiment tags

[0851] Output: Text displayed on the screen, translation results played back aloud, and emotional information displayed

[0852] Step 8:

[0853] Checked and supplemented by experts

[0854] User (medical professional): Checks the translated content and the recognized sentiment, and provides corrections or additional information as needed. Specifically, the professional reviews the displayed translation, makes corrections if necessary, and enters a follow-up question. For example, a doctor might enter a follow-up question such as, "What causes headaches?"

[0855] Input: Translated text data and recognized sentiment tags

[0856] Output: Corrected translation and additional questions

[0857] Step 9:

[0858] Data encryption and protection

[0859] Server: All data is protected with advanced encryption technology. Specifically, voice and text data is encrypted with the AES encryption algorithm, and data in transit is securely transmitted and received using the TLS protocol.

[0860] Input: Audio data, text data, emotion data

[0861] Output: AES encrypted data, secure communication via TLS protocol

[0862] Step 10:

[0863] Training a generative AI model

[0864] Server: Trains the AI ​​model to understand cultural differences and multilingual environments. Specifically, the AI ​​model learns using multilingual datasets to improve its ability to understand Japanese nuances and cultural background.

[0865] Input: Multilingual dataset

[0866] Output: Generative AI models that are culturally and multilingually responsive

[0867] (Application example 2)

[0868] 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."

[0869] Elderly and foreign customers can sometimes find it difficult to communicate smoothly with store staff. In such situations, customers may not receive appropriate service due to language barriers or a lack of understanding of their emotions. Particularly in physical stores, where appropriate responses are required in real time, a system that solves these issues is necessary.

[0870] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generation AI model, means for displaying the translated text and emotional state, and means for audibly playing back the translated text and emotional state. This makes it possible to translate voice data into multiple languages ​​and display and play back the emotional state in real time, allowing elderly people and foreign customers to receive appropriate services.

[0871] "Voice data" is a digital recording of a user's speech.

[0872] "Collection means" refers to the devices and techniques used to collect audio data.

[0873] "Preprocessing" refers to performing data processing such as noise removal and normalization on collected voice data.

[0874] A "speech recognition model" is an algorithm or machine learning model that analyzes voice data and converts it into text data.

[0875] "Text" is character data converted from audio data.

[0876] A "multilingual generative AI model" is an artificial intelligence model used to translate text input in one language into another language.

[0877] "Translation" is the conversion of text written in one language into another.

[0878] "Display means" refers to devices or techniques for visually presenting the translated text and emotional state to the user.

[0879] "Audio playback" refers to synthesizing text data into audio and providing it to the user audibly.

[0880] The "emotional state" indicates the emotional state recognized from the user's speech and facial expressions.

[0881] "Means of expert checking and supplementation" refers to a method in which an expert checks the accuracy of the translation results and provides corrections or additional information as necessary.

[0882] "Encryption" is a technology that converts data based on a specific algorithm to prevent unauthorized access by third parties.

[0883] "Cultural training" is the process by which generative AI models learn to adapt to multicultural environments.

[0884] This invention is a system for facilitating communication between elderly and foreign customers and store staff in brick-and-mortar stores. This system collects user speech data and performs preprocessing, translation, and emotion recognition using the following means:

[0885] 1. Audio data collection:

[0886] The server uses a terminal to collect voice data, such as a microphone built into smart glasses or a smartphone, which allows the server to capture customer speech in real time.

[0887] 2. Preprocessing the audio data:

[0888] The server then denoises and normalizes the received audio data using voice processing APIs from Amazon AWS and Google Cloud, making the data easier to analyze.

[0889] 3. Audio to text conversion:

[0890] The server then feeds the preprocessed audio data into a speech recognition model and converts it into text using an automatic speech recognition algorithm such as IBM Watson Speech to Text.

[0891] 4. Text translation:

[0892] The server translates the converted text using a multilingual generative AI model, using a translation engine such as the DeepL translation API, enabling smooth communication between different languages.

[0893] 5. Emotion recognition:

[0894] The server uses a natural language processing API (NLP) from Google Cloud to analyze the user's emotions from the text data. By identifying the user's emotional state, it can provide a more appropriate response.

[0895] 6. Displaying translation results and emotional state and playing audio:

[0896] The translated text and emotional state are displayed on the HUD (heads-up display) or smartphone screen, and the translation is played back via speech synthesis for real-time understanding by both the customer and store staff.

[0897] 7. Expert check and completion:

[0898] Experts check the accuracy of the translated content and provide corrections or additional information as needed, which increases the reliability of the system.

[0899] For example, when a store clerk asks, "Welcome. What can I do for you?", an elderly customer responds, "My feet have been hurting recently, so it's a bit difficult to walk." This speech data is collected, preprocessed, and converted into text using a speech recognition model. Next, this text is translated into English using the DeepL translation API, and the emotional state is identified as "negative" using Google Cloud's Natural Language API. Finally, the HUD displays, "My feet have been hurting recently, so it's a bit difficult to walk. Emotion score: -0.3," and the translation result is conveyed to the store clerk through audio playback.

[0900] An example of a prompt sentence for the generative AI model is as follows:

[0901] "Audio of a patient speaking Japanese in a hospital"

[0902] "output_language: English"

[0903] "output_emotion: emotion analysis"

[0904] "expected_output: Translation result: 'My feet have been hurting recently, so it's a bit difficult to walk.' Sentiment score: '-0.3'"

[0905] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0906] Step 1:

[0907] The user speaks into the microphone built into the smart glasses or smartphone. The device captures the customer's voice data and transmits it to the server in real time. The input is the customer's voice data, and the output is the digital voice data sent to the server.

[0908] Step 2:

[0909] The server denoises and normalises the received audio data. This process uses the audio processing APIs of Amazon AWS and Google Cloud. The input is the received audio data, and the output is the denoised and normalised audio data.

[0910] Step 3:

[0911] The server inputs the preprocessed audio data into a speech recognition model and converts it into text. IBM Watson Speech to Text is used here. The input is the preprocessed audio data, and the output is text data.

[0912] Step 4:

[0913] The server translates the converted text using a multilingual generative AI model. This process uses the DeepL translation API. The input is text data, and the output is the translated text.

[0914] Step 5:

[0915] The server uses a natural language processing API to analyze the user's emotional state from the translated text. Google Cloud's Natural Language API is used here. The input is the translated text data, and the output is the emotion score and emotion type.

[0916] Step 6:

[0917] The server displays the translated text and emotional state on a HUD (Heads-Up Display) or smartphone screen. The displayed content includes the translated text and the emotional score. The input is the translated text and the emotional state, and the output is the displayed information.

[0918] Step 7:

[0919] The server synthesizes the translation result and plays it back, allowing the user (store clerk) to listen to the translation in real time. The input is the translated text, and the output is audio data.

[0920] Step 8:

[0921] The expert (clerk) checks the displayed and played translation results and emotional state, and provides corrections or additional information as needed. The input is the translated text and emotional state, and the output is the corrected or supplemented information.

[0922] 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.

[0923] 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.

[0924] 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.

[0925] [Third embodiment]

[0926] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0927] 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.

[0928] 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).

[0929] 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.

[0930] 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.

[0931] 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).

[0932] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0933] 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.

[0934] 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.

[0935] 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.

[0936] 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.

[0937] 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."

[0938] This is a system to support smooth communication between elderly and foreign patients and medical professionals. The system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment.

[0939] Program implementation and processing explanation

[0940] Basic configuration

[0941] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[0942] Overview of the main measures that will emerge

[0943] 1. Collection of audio data

[0944] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[0945] 2. Preprocessing of audio data

[0946] Server: Performs noise reduction and normalization on the received audio data, removing unnecessary noise and converting the audio data into a format that is easier to analyze.

[0947] 3. Speech to text conversion

[0948] Server: The preprocessed speech data is analyzed using a speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[0949] 4. Text Translation

[0950] Server: The converted text data is translated in real time using a multilingual generative AI model. For example, the Japanese text "I have a headache" is translated into English as "I have a headache."

[0951] 5. Display and playback of translated text and audio

[0952] On your device: The translated text will be displayed on your device screen and played back using speech synthesis, if available. For example, the translation will be displayed on your device and the speech synthesis will say "I have a headache."

[0953] 6. Expert check and completion

[0954] User (medical professional): Checks the translation results and corrects or completes them as necessary. This improves the accuracy and reliability of the translation. For example, a medical professional corrects "windpipe" to "trachea."

[0955] 7. Data Encryption and Protection

[0956] Server: All data is protected with advanced encryption technology, for example, using the AES encryption algorithm, and data in transit is sent and received over secure protocols (TLS / SSL).

[0957] 8. Training a generative AI model

[0958] Server: Generative AI models are trained to understand cultural differences and sensitivities, for example, using datasets specific to Japanese customs and expressions.

[0959] Specific examples

[0960] Consider a scenario in which a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital.

[0961] 1. Voice input: The patient speaks to the terminal in English, "I have a headache."

[0962] 2. Data preprocessing and speech recognition: This speech data is sent to the server, where it is noise-reduced and converted to text using an ASR model.

[0963] 3. Text translation: A generative AI model translates the English phrase "I have a headache" into Japanese phrase "I have a headache."

[0964] 4. Display and audio output: The translated Japanese text is displayed on the doctor's device and played as audio if necessary.

[0965] 5. Expert review: A doctor reviews the translation and asks additional questions or makes a diagnosis.

[0966] The above steps will enable rapid and accurate medical communication that transcends language barriers.

[0967] The processing flow will be explained below.

[0968] Step 1:

[0969] Audio data collection

[0970] Device: The patient or healthcare professional speaks into the device. The device's built-in microphone captures the audio and begins recording.

[0971] Example of operation: When a patient says "I have a headache" to the device, the voice is recorded.

[0972] Step 2:

[0973] Sending audio data

[0974] Terminal: Streams recorded audio data to the server in real time.

[0975] How it works: Captured audio data is converted into digital form and sent over a network to a server.

[0976] Step 3:

[0977] Audio data preprocessing

[0978] Server: Performs noise reduction and normalization on the received audio data, converting it into a clear and consistent format.

[0979] How it works: A filtering algorithm is applied to remove background noise and even out volume variations.

[0980] Step 4:

[0981] Text conversion using speech recognition models

[0982] Server: The preprocessed speech data is input into the speech recognition model and converted into text.

[0983] Example of how it works: A speech recognition algorithm analyzes the characteristics of a voice and converts the speech "I have a headache" into the text "I have a headache."

[0984] Step 5:

[0985] Text translation

[0986] Server: Using a generative AI model, translate text data into the target language. Use a multilingual translation engine.

[0987] Example of how it works: The text "I have a headache" is translated into English as "I have a headache".

[0988] Step 6:

[0989] Display and playback of translation results

[0990] On Device: Displays translated text and audio on the user's device, optionally using speech synthesis to play the translation aloud.

[0991] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache" in English.

[0992] Step 7:

[0993] Checked and supplemented by experts

[0994] User (medical professional): Review the translated content and provide corrections or additional information as needed.

[0995] Example of how it works: A doctor checks the translation results and corrects any inaccuracies. For example, he or she asks an additional question in Japanese: "Please tell me the cause of my headache."

[0996] Step 8:

[0997] Data encryption and protection

[0998] Server: We encrypt all data sent and received and use advanced security protocols to prevent unauthorized access to your data by third parties.

[0999] How it works: The server encrypts data using the AES encryption algorithm, and the encrypted data is sent and received via the TLS protocol.

[1000] Step 9:

[1001] Training a generative AI model

[1002] Server: Trains AI models to understand cultural differences and multilingual environments, thereby reducing mistranslations and cultural misunderstandings.

[1003] How it works: An AI model trains on a server using a multilingual dataset to improve its ability to understand Japanese-specific nuances.

[1004] Example 1

[1005] 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."

[1006] In today's medical settings, communication barriers between elderly or foreign patients and medical professionals are a serious problem. It is particularly difficult to communicate smoothly with patients who have language differences, hearing impairments, or cultural backgrounds, which can delay appropriate diagnoses and treatment. A system that can resolve these issues and support accurate and prompt communication in medical settings is needed.

[1007] 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.

[1008] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for experts to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for noise reduction and normalization, means for inputting voice into a voice recognition model, and means for synthesizing voice, thereby enabling accurate and prompt communication between patients and medical professionals with different languages, hearing impairments, or cultural backgrounds.

[1009] An "audio data collection means" is a device or method for capturing and transmitting audio from a patient or medical professional to a server.

[1010] The "audio data preprocessing means" refers to a device or method for converting collected audio data into a form that is easy to analyze by performing processes such as noise reduction and normalization.

[1011] A "speech recognition model" is a machine learning algorithm that analyzes preprocessed speech data and converts it into text data.

[1012] A "multilingual generative AI model" is an artificial intelligence model used to translate between different languages ​​in real time.

[1013] "Display and playback means" refers to a device or method for displaying the translated text on the screen of the terminal and, if necessary, synthesizing it to speech and playing it back.

[1014] "Expert checking and completion means" refers to an interface and method for medical professionals to check the translation results and make corrections or completions as necessary.

[1015] A "data encryption means" is a device or method for encrypting data using an encryption algorithm to protect all data processed within the system.

[1016] A "means for training a generative AI model" is a device or method for training a generative AI model with a particular dataset so that the model understands cultural differences and sensitivities.

[1017] "Noise reduction and normalization means" refers to a device or method for reducing noise in audio data during pre-processing and adjusting the volume and pitch to a consistent level.

[1018] A "means for inputting speech into a speech recognition model" is a device or method for appropriately inputting preprocessed speech data into a speech recognition model.

[1019] "Speech synthesis means" refers to a device or method for generating speech based on text data and reproducing it.

[1020] This invention relates to a system that supports smooth communication between elderly and foreign patients and medical professionals. Specifically, it enables accurate and prompt medical treatment by providing voice, text, translation, and real-time medical information.

[1021] Fundamental System Configuration

[1022] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[1023] Technical Details

[1024] Hardware and Basic Operation

[1025] The device is used by the patient or a medical professional and collects voice data through a built-in microphone. The device can be a smartphone, tablet, or dedicated voice collection device. The collected voice data is sent to a server in real time. For example, if a patient says, "I have a headache," the voice is captured instantly.

[1026] The server preprocesses the received audio data. The main processes are noise reduction and normalization. Noise reduction removes background noise, and normalization maintains the quality of the audio data.

[1027] The preprocessed speech data is then converted into text data using an Automatic Speech Recognition (ASR) model. For example, the Japanese phrase "I have a headache" is converted directly into text.

[1028] The converted text is then translated in real time by a multilingual generative AI model, pre-trained to understand cultural differences and sensitivities, for example, translating the Japanese text "Watashi wa tou ga kaiyuu" (I have a headache) into English.

[1029] The translated text and audio are then displayed and played back on the device. The translated text is displayed on the device screen and the translation is played back using speech synthesis. For example, the translated "I have a headache" is displayed and played back as audio.

[1030] The user (medical professional) can check the translation results through the interface provided by the system and make corrections or additions as necessary, thereby improving the accuracy and reliability of the translation. For example, a medical professional can correct "windpipe" to "trachea."

[1031] The server protects all data with advanced encryption technology, specifically using the AES encryption algorithm, and data in transit is sent and received via secure protocols (TLS / SSL).

[1032] Specific examples

[1033] Consider a case where a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital. The following steps are taken:

[1034] 1. Voice input:

[1035] The patient speaks to the terminal in English, "I have a headache."

[1036] 2. Data preprocessing and speech recognition:

[1037] This audio data is sent to a server, where it is noise reduced and normalized before being converted to text using an ASR model.

[1038] 3. Text Translation:

[1039] The server uses a generative AI model to translate the English phrase "I have a headache" into Japanese phrase "I have a headache."

[1040] 4. Display and audio output:

[1041] The translated Japanese text is displayed on the doctor's device and the audio is played back as "I have a headache."

[1042] 5. Expert review:

[1043] The doctor will check the translated content and provide appropriate diagnosis and treatment.

[1044] Prompt Sentence Examples

[1045] Examples of prompts to enter into this system are:

[1046] "If a patient says, 'I have a headache,' in English, convert that audio data into Japanese text and speech for the doctor to review."

[1047] The above procedures and system configuration will enable rapid and accurate medical communication that transcends language barriers.

[1048] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1049] Step 1:

[1050] Audio data collection

[1051] The device captures the voice of the patient or medical professional. When the patient speaks into the device, the built-in microphone collects the voice and transmits it to the server in real time. The input is voice data, and the output is raw voice data sent to the server.

[1052] Specific behavior:

[1053] The patient speaks into the terminal's microphone, "I have a headache."

[1054] The device collects the voice data and sends it to the server.

[1055] Step 2:

[1056] Audio data preprocessing

[1057] The server performs noise reduction and normalization on the received audio data, converting it into a form that is easier to analyze. The input is raw audio data, and the output is preprocessed audio data.

[1058] Specific behavior:

[1059] The server receives the audio data.

[1060] Apply a noise reduction algorithm to remove background noise.

[1061] Normalizes the volume and pitch of audio data.

[1062] Step 3:

[1063] Speech-to-text conversion

[1064] The server inputs the preprocessed speech data into an Automatic Speech Recognition (ASR) model and converts it into text data. The input is the preprocessed speech data, and the output is text data.

[1065] Specific behavior:

[1066] The server inputs the speech data into the ASR model.

[1067] The ASR model analyzes the speech data and generates the text "I have a headache."

[1068] Step 4:

[1069] Text translation

[1070] The server translates the converted text using a generative AI model. The input is text data, and the output is translated text data.

[1071] Specific behavior:

[1072] The server inputs text into the generative AI model.

[1073] A generative AI model translates the Japanese text "I have a headache" into English "I have a headache."

[1074] Step 5:

[1075] View and play translated text and audio

[1076] The terminal displays the translated text received from the server on the screen and plays it back using speech synthesis if necessary. The input is the translated text data, and the output is the displayed text and speech data.

[1077] Specific behavior:

[1078] The terminal receives the translation result from the server.

[1079] The translated text is displayed on the screen.

[1080] Uses a speech synthesis engine to play the translated text aloud.

[1081] Step 6:

[1082] Checked and supplemented by experts

[1083] The user (medical professional) checks the translation results displayed on the terminal and makes corrections or completions as necessary. The input is the translated text data, and the output is the corrected or completed text data.

[1084] Specific behavior:

[1085] The medical professional checks the translation results displayed on the device's display.

[1086] If the translation is not accurate, correct it.

[1087] Step 7:

[1088] Data encryption and protection

[1089] The server encrypts all audio and text data. The input is the processed data and the output is the encrypted data.

[1090] Specific behavior:

[1091] The server encrypts the audio and text data using the AES encryption algorithm.

[1092] Encrypted data is sent and received using secure protocols (TLS / SSL).

[1093] Step 8:

[1094] Training a generative AI model

[1095] The server trains the generative AI model to understand cultural differences and sensitivities: the input is a specific culture and language dataset, and the output is an optimized generative AI model.

[1096] Specific behavior:

[1097] The server inputs specific culture and language datasets into the AI ​​model.

[1098] The server optimizes the parameters of the AI ​​model and performs training.

[1099] (Application example 1)

[1100] 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."

[1101] To ensure smooth communication between elderly and foreign patients and medical professionals, accurate real-time speech recognition and translation, as well as secure processing and transmission of the data, are necessary. However, current systems lack the ability to process voice data in real time, securely encrypt the data, and train generative AI models that take cultural differences into account, hindering accurate and prompt medical responses.

[1102] 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.

[1103] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a speech recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for an expert to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for securely storing and transmitting the encrypted data, and means for processing the voice data collected by the terminal in real time, thereby enabling real-time and accurate medical communication between patients and medical professionals.

[1104] "Audio data" refers to data in which an audio signal is recorded as digital information.

[1105] A "collection means" is a device or system for acquiring and transmitting audio data to another processing unit.

[1106] The "preprocessing means" is a device or system that performs preprocessing such as noise removal and normalization on collected voice data.

[1107] A "voice recognition model" is an algorithm or AI model that analyzes voice data and converts it into a string of characters (text).

[1108] "Text" is digital data expressed as character information.

[1109] A "multilingual generative AI model" is an AI model that supports multiple languages ​​and translates from one language to another in real time.

[1110] A "translation tool" is a device or system that converts text in a particular language into another language.

[1111] "Display means" refers to a device or interface for visually showing the translation results to the user.

[1112] "Playback means" refers to a device or system that provides the translation result to the user as audio.

[1113] "Means for expert checking and correction" refers to a device or system that allows an expert to verify the accuracy of the translation results and correct them if necessary.

[1114] An "encryption means" is a device or system that encrypts data to protect the security of the data.

[1115] A "training means for a generative AI model to take cultural differences into account" is a device or system that trains a generative AI model to understand differences in cultural background and expression.

[1116] A "real-time processing means" is a device or system that instantly analyzes and processes audio data and provides results to the user in real time.

[1117] "Storage and transmission means" means a device or system that securely stores encrypted data and transmits it wherever needed.

[1118] To implement this invention, a system must be constructed by combining a series of hardware and software. The specific system configuration, processing procedures, and usage methods are described below.

[1119] System configuration

[1120] This system is a multi-function system for speech recognition, real-time translation, data encryption, and displaying and playing back the results.

[1121] Hardware used

[1122] 1. Collection device: A device for collecting voice data, such as a smartphone or head-mounted display. This device has a built-in highly sensitive microphone.

[1123] 2. Servers: High-performance servers are used for data pre-processing, speech recognition, text conversion, translation, and data encryption, including cloud servers.

[1124] 3. Display device: A device that displays the translated data and plays it back as audio. This includes smartphones and tablets.

[1125] Software used

[1126] 1. Speech recognition library: A library for converting voice data to text (e.g., speech_recognition).

[1127] 2. Translation library: A library for translating text into another language (e.g. googletrans).

[1128] 3. Cryptography library: A library for performing data encryption (e.g., the Fernet module for cryptography).

[1129] System processing explanation in natural language

[1130] The server receives voice data from a collection terminal, which is used by a patient or a medical professional to capture the voice data. The captured voice data undergoes pre-processing for noise reduction and normalization before being sent to the server.

[1131] The server converts the preprocessed speech data into text using a speech recognition library. This text data is translated into the specified language in real time using a multilingual generative AI model. For example, if a patient says "I have a headache" in English, this will be translated into Japanese as "I have a headache."

[1132] The translated text is sent to a display terminal where it is displayed on the screen and played back as audio. The system also includes an interface that allows experts to check the accuracy of the translation and make corrections or additions if necessary.

[1133] All data is encrypted using advanced encryption technology and transmitted over secure protocols. The server uses the AES encryption algorithm and TLS / SSL protocols to ensure data security.

[1134] Examples of concrete examples and prompts

[1135] Specific usage scenarios include the following:

[1136] 1. Example: When an English-speaking patient visits a Japanese hospital, they use their smartphone to communicate their symptoms to the doctor. When they speak into their smartphone, for example, "I have a headache," this voice data is translated into Japanese and displayed and played on the doctor's device.

[1137] 2. Example prompt:

[1138] I'm an English-speaking patient with a headache. I say "I have a headache" in English to my smartphone. Please translate the voice data into Japanese and display it so that my doctor can see it. Also, please encrypt the data to ensure its security.

[1139] This invention enables real-time and accurate medical communication between elderly and foreign patients and medical professionals.

[1140] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1141] Step 1:

[1142] The collection terminal captures the speech data uttered by the patient or medical professional. The input is the speech signal collected from the microphone of the collection terminal. This speech signal is converted into a digital format and subjected to noise reduction and normalization processing. The output is the pre-processed speech data.

[1143] Step 2:

[1144] The preprocessed voice data is sent from the collection terminal to the server. The input is the preprocessed voice data. The collection terminal sends the data to the server using a transmission protocol. The output is the voice data received at the server side.

[1145] Step 3:

[1146] The server performs speech recognition processing on the received audio data. The input is noise-reduced and normalized audio data. The server uses a speech recognition library (e.g., speech_recognition) to convert the audio data into text data. The output is text data.

[1147] Step 4:

[1148] The server translates the converted text data using a multilingual generative AI model. The input is text data generated by a speech recognition model. A translation library (e.g., GoogleTrans) is used to translate the text into the specified language in real time. The output is the translated text data.

[1149] Step 5:

[1150] The server encrypts the translated text data. The input is the translated text data. The data is encrypted using an encryption library (e.g. Fernet). The output is the encrypted text data.

[1151] Step 6:

[1152] Encrypted text data is sent from the server to the display terminal. The input is the encrypted text data. The server transfers the data using a secure protocol (e.g., TLS / SSL). The output is the encrypted text data received by the display terminal.

[1153] Step 7:

[1154] The display terminal decrypts encrypted text data received from the server. The input is the encrypted text data. The display terminal uses the encryption library to decrypt the data. The output is the decrypted text data.

[1155] Step 8:

[1156] The display terminal displays and plays back the decoded text data. The input is the decoded text data. The user interface displays the text data on the screen and uses speech synthesis to play the text aloud. The output is a translation result that can be recognized visually and audibly by the user.

[1157] Step 9:

[1158] The expert user uses a display terminal to check the accuracy of the translated text data and make corrections or completions as necessary. The input is the decoded and displayed text data. The user makes corrections through the interface, and the results are reflected in the system. The output is the corrected and completed accurate translation data.

[1159] This series of steps enables accurate and secure medical communication in real time.

[1160] 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.

[1161] This invention is a system for supporting smooth communication and emotion recognition between elderly and foreign patients and medical professionals. This system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it is possible to make diagnoses and treatments that take into account the patient's emotional state.

[1162] Program implementation and processing explanation

[1163] Basic configuration

[1164] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, and a terminal that displays and plays back the converted and translated information. An interface is also provided for experts to check and complete the translation results and recognized emotions. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[1165] Overview of the main measures that will emerge

[1166] 1. Collection of audio data

[1167] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[1168] 2. Sending audio data

[1169] Terminal: Streams recorded audio data to the server in real time. The captured audio data is converted into a digital format and sent over the network to the server.

[1170] 3. Preprocessing of audio data

[1171] Server: Noise reduction and normalization processes are performed on the received audio data, converting the audio into a clear and consistent format. Filtering algorithms are applied to remove background noise and even out volume variations.

[1172] 4. Emotion Recognition by Emotion Engine

[1173] Server: Analyzes the preprocessed audio and video data to recognize the user's emotional state. For example, it identifies emotions such as "anger," "sadness," and "surprise." The emotion recognition algorithm reads the user's emotions from their voice intonation and facial expressions.

[1174] 5. Speech to Text Conversion

[1175] Server: The preprocessed speech data is input into the speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[1176] 6. Text Translation

[1177] Server: Translates text data into the target language using a generative AI model. It uses a multilingual translation engine. For example, the text "I have a headache" is translated into English as "I have a headache."

[1178] 7. Display and playback of translation results and emotions

[1179] Terminal: The translated text and recognized emotion information are displayed on the user's terminal. If necessary, the translation result is played back aloud using speech synthesis. For example, the translation result is displayed on the doctor's terminal, and the speech synthesis reproduces "I have a headache" in English. The display screen also indicates the patient's emotional state (e.g., "anxiety").

[1180] 8. Expert check and completion

[1181] User (medical professional): Review the translated content and the perceived sentiment, and provide corrections or additional information as needed. This improves the accuracy and reliability of the translation. For example, a doctor may review the translation and ask a follow-up question such as "What is causing your headache?", as well as provide suggestions to ease the patient's anxiety.

[1182] 9. Data Encryption and Protection

[1183] Server: All data is protected with advanced encryption technology. Voice and text data is encrypted with the AES encryption algorithm, and data in transit is also securely transmitted via the TLS protocol.

[1184] 10. Training a generative AI model

[1185] Server: Trains AI models to understand cultural differences and multilingual environments. Using multilingual datasets, AI models learn and improve their ability to understand Japanese-specific nuances.

[1186] These steps not only enable fast and accurate medical communication across language barriers, but also provide a more humane response that takes into account the patient's emotional state.

[1187] The processing flow will be explained below.

[1188] Step 1:

[1189] Audio data collection

[1190] Device: The patient or healthcare professional speaks into the device, which has a built-in microphone that captures the voice and records the data in real time.

[1191] Example of operation: When a patient says "I have a headache" to the device, the voice is instantly recorded and saved as digital audio data.

[1192] Step 2:

[1193] Sending audio data

[1194] Terminal: Recorded audio data is sent to the server in streaming format.

[1195] How it works: Send captured audio data to a server using any Wi-Fi or 4G network within range.

[1196] Step 3:

[1197] Audio data preprocessing

[1198] Server: Performs noise reduction and normalization on the received audio data, making it easier to analyze.

[1199] How it works: Using a noise filtering algorithm on the server side to remove background noise and enhance the audio signal.

[1200] Step 4:

[1201] Emotion recognition by emotion engine

[1202] Server: Analyzes the pre-processed audio and video data (if necessary) and recognizes the user's emotions. An emotion recognition algorithm is used.

[1203] Example of how it works: The server analyzes the intonation, speed, and emphasized parts of the voice to identify emotions such as "anger," "sadness," and "surprise." It also uses facial expression data recognized using a camera.

[1204] Step 5:

[1205] Speech-to-text conversion

[1206] Server: The preprocessed speech data is input into the speech recognition model (ASR) and converted into corresponding text data.

[1207] Example of how it works: A speech recognition algorithm listens for speech characteristics and converts the speech "I have a headache" into text "I have a headache."

[1208] Step 6:

[1209] Text translation

[1210] Server: Using a generative AI model, the converted text data is translated into the required language using a multilingual translation engine.

[1211] Working example: Translate the Japanese phrase "I have a headache" into English.

[1212] Step 7:

[1213] Display and playback of translation results and emotional information

[1214] Device: The translated text and recognized emotion information are displayed on the user's device, and the translation result is played aloud using speech synthesis if necessary.

[1215] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache." Emotional information such as "Emotion: Anxiety" is also displayed on the display screen.

[1216] Step 8:

[1217] Checked and supplemented by experts

[1218] User (medical professional): Checks the content of the translated text and sentiment information, and makes corrections or asks additional questions as needed, thereby improving the accuracy of the diagnosis.

[1219] Example of how it works: A doctor checks the translation results and emotional information, and then asks an additional question in Japanese, such as "Please tell me the cause of my headache." The system translates this into English in real time.

[1220] Step 9:

[1221] Data encryption

[1222] Server: All data sent and received is protected with the AES encryption algorithm, and communication is carried out via the TLS / SSL protocol.

[1223] Example of operation: All data, including recorded voice data, translated text, and emotional information, is encrypted and transmitted over a secure communication channel.

[1224] Step 10:

[1225] Training a generative AI model

[1226] Server: Trains generative AI models to understand cultural differences and multilingual environments, using multilingual datasets for training.

[1227] How it works: The server trains an AI model using a dataset on Japanese culture to understand Japanese-specific expressions and nuances.

[1228] The above processing flow enables accurate and prompt communication between patients and medical professionals, and also enables responses that take into account the patient's emotional state.

[1229] Example 2

[1230] 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."

[1231] It is difficult for elderly and foreign patients to communicate smoothly with medical professionals, especially when it comes to accurately recognizing emotions and providing appropriate responses. Furthermore, in multilingual medical communication, it is necessary to ensure the accuracy and reliability of translations. Furthermore, the security of medical information is also a major issue.

[1232] 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.

[1233] In this invention, the server includes a means for transmitting voice data in real time, a means for performing noise reduction and normalization on the preprocessed voice data, and a means for analyzing the preprocessed voice data to recognize emotions. This enables medical communication that takes into account the user's emotions while overcoming language barriers. It also improves the accuracy and reliability of translations and ensures the security of medical information.

[1234] "Means for collecting voice data" refers to the function of capturing voices spoken by patients or medical professionals using the device's microphone.

[1235] The "means for transmitting audio data to a server in real time" refers to a mechanism for converting captured audio data into a digital format in real time and transmitting it to a server via a network.

[1236] The "noise reduction and normalization means" is a function that removes background noise from the received audio data and performs processing to equalize the volume.

[1237] "Means for recognizing emotions" refers to a technology that analyzes pre-processed voice data and identifies the user's emotional state from the intonation and tempo of the voice.

[1238] A "speech recognition model" is a technology that analyzes voice data and converts it into text format, often using automatic speech recognition (ASR) algorithms.

[1239] A "multilingual generative AI model" is an AI technology that supports multiple languages ​​and translates input text data into other languages.

[1240] "Display and playback means" refers to the functionality of providing the translated text and emotional information to the user visually and audibly.

[1241] "Means for expert checking and correction" refers to a system in which experts can check and correct the content of translated text and the results of emotion recognition.

[1242] "Means for encrypting data" refers to a method of protecting all data with advanced encryption technology, ensuring that data is sent and received securely during communication.

[1243] "Training measures to take cultural differences into account" refers to the learning process that generative AI models undergo to understand the cultural nuances of each language and translate or respond appropriately.

[1244] This invention is a system that enables smooth communication between elderly and foreign patients and medical professionals, and also enables emotion recognition. This system provides voice, text, translation, and real-time medical information to support accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it becomes possible to diagnose and treat patients taking into account their emotional state.

[1245] Basic configuration

[1246] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, a terminal that displays and plays back the converted and translated information, and an interface for experts to check and complete the translation results and recognized emotions. The system also employs data encryption and confidentiality technologies to meet high security standards.

[1247] Hardware and Software

[1248] 1. Device:

[1249] It is a device with a user interface and a built-in microphone for collecting and capturing voice data. The device also has a processor, memory, and a network module, and transmits the voice data to a server in real time.

[1250] 2. Server:

[1251] The server preprocesses the received voice data and performs emotion and speech recognition. It has a high-performance processor, memory, and data storage, and runs speech recognition models and generative AI models, which then perform noise reduction and data analysis.

[1252] 3. Software:

[1253] The speech recognition model running on the server uses an automatic speech recognition (ASR) algorithm to convert preprocessed speech data into text. The generative AI model is multilingual and translates text data into other languages. In particular, it employs generative AI technology that has a proven track record as a translation engine.

[1254] 4. Emotion Engine:

[1255] This engine identifies the user's emotions by analyzing the intonation and waveform of voice data, and recognizes a variety of emotional states (e.g., anger, sadness, anxiety).

[1256] 5. Security Technology:

[1257] The AES encryption algorithm is used to encrypt data, and data is sent and received securely via the TLS protocol during communication.

[1258] 6. Expert interface:

[1259] This interface allows experts to review the translation results and the perceived sentiment, and provide corrections or additional information as needed. Users can input appropriate actions, thereby improving the accuracy and reliability of the translation.

[1260] Specific examples

[1261] For a specific scenario, use the following prompt:

[1262] Prompt Sentence Examples

[1263] The patient is expected to say to the device in Japanese, "I have chest pain," and the doctor will understand this in English.

[1264] The device captures the patient's voice saying "I feel chest pain," converts it into digital format, and sends it to the server. The server performs noise reduction on the received voice data and then uses an emotion engine to recognize the emotion "anxiety." Next, a speech recognition model converts the voice into text, generating the text data "I feel chest pain." This text data is translated into English by a generative AI model, resulting in the result "I feel chest pain." The translation result and emotion information are displayed on the device and, if necessary, played back using a speech synthesis function. Finally, an expert reviews the result and makes corrections or asks additional questions as necessary. The expert can then ask an additional question, "When did the chest pain start?" and provide measures to alleviate the patient's anxiety.

[1265] As described above, this system not only enables fast and accurate medical communication across language barriers, but also provides a humane response that takes into account the patient's emotional state.

[1266] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1267] Step 1:

[1268] Audio data collection

[1269] Device: The patient or medical professional speaks into the device's microphone. Voice data is input in analog format, and the microphone captures the voice and converts it into digital format. Specifically, the device activates the microphone and records the patient saying, "I have a headache." The recorded voice data is stored in a buffer.

[1270] Input: Patient speech (e.g., "I have a headache")

[1271] Output: Digital audio data

[1272] Step 2:

[1273] Sending audio data

[1274] Terminal: Converts recorded audio data into digital format in real time and streams it to the server over the network. Specifically, the terminal packs audio data into packets and sends them to the server via the network module.

[1275] Input: Recorded digital audio data

[1276] Output: Audio data sent to the server

[1277] Step 3:

[1278] Audio data preprocessing

[1279] Server: Performs noise reduction and normalization on the received audio data to convert it into a clear, uniform format. Specifically, the server filters the audio data, removes background noise, and normalizes volume fluctuations.

[1280] Input: Audio data sent to the server

[1281] Output: Noise-reduced and normalized audio data

[1282] Step 4:

[1283] Emotion recognition by emotion engine

[1284] Server: Analyzes the preprocessed voice data and identifies the user's emotions from the intonation and waveform of the voice data. Specifically, the emotion engine analyzes the frequency components of the voice data and recognizes emotions such as "anger," "sadness," and "surprise."

[1285] Input: Preprocessed audio data

[1286] Output: Recognized emotion tag (e.g., "anxiety")

[1287] Step 5:

[1288] Speech-to-text conversion

[1289] Server: Using a speech recognition model, the preprocessed speech data is converted into text format. Specifically, the speech data is broken down into phonemes, which are then reconstructed into strings of characters using a language model to generate text data.

[1290] Input: Preprocessed audio data

[1291] Output: Text data (e.g., "I have a headache")

[1292] Step 6:

[1293] Text translation

[1294] Server: Translates text data into other languages ​​using a generative AI model. Specifically, the generative AI model first analyzes the input text and then generates a translation based on the grammar and vocabulary of the target language. For example, "I have a headache" is translated into "I have a headache."

[1295] Input: Text data (e.g., "I have a headache")

[1296] Output: Translated text data (e.g., "I have a headache")

[1297] Step 7:

[1298] Display and playback of translation results and emotions

[1299] Terminal: The translated text and emotion information are displayed on the user's terminal and played back using a speech synthesis function as needed. Specifically, the terminal displays the text on the screen and uses a speech synthesis engine to play back the translation result aloud. For example, the translation result is displayed on the doctor's terminal and the phrase "I have a headache" is played back.

[1300] Input: Translated text data and recognized sentiment tags

[1301] Output: Text displayed on the screen, translation results played back aloud, and emotional information displayed

[1302] Step 8:

[1303] Checked and supplemented by experts

[1304] User (medical professional): Checks the translated content and the recognized sentiment, and provides corrections or additional information as needed. Specifically, the professional reviews the displayed translation, makes corrections if necessary, and enters a follow-up question. For example, a doctor might enter a follow-up question such as, "What causes headaches?"

[1305] Input: Translated text data and recognized sentiment tags

[1306] Output: Corrected translation and additional questions

[1307] Step 9:

[1308] Data encryption and protection

[1309] Server: All data is protected with advanced encryption technology. Specifically, voice and text data is encrypted with the AES encryption algorithm, and data in transit is securely transmitted and received using the TLS protocol.

[1310] Input: Audio data, text data, emotion data

[1311] Output: AES encrypted data, secure communication via TLS protocol

[1312] Step 10:

[1313] Training a generative AI model

[1314] Server: Trains the AI ​​model to understand cultural differences and multilingual environments. Specifically, the AI ​​model learns using multilingual datasets to improve its ability to understand Japanese nuances and cultural background.

[1315] Input: Multilingual dataset

[1316] Output: Generative AI models that are culturally and multilingually responsive

[1317] (Application example 2)

[1318] 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."

[1319] Elderly and foreign customers can sometimes find it difficult to communicate smoothly with store staff. In such situations, customers may not receive appropriate service due to language barriers or a lack of understanding of their emotions. Particularly in physical stores, where appropriate responses are required in real time, a system that solves these issues is necessary.

[1320] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generation AI model, means for displaying the translated text and emotional state, and means for audibly playing back the translated text and emotional state. This makes it possible to translate voice data into multiple languages ​​and display and play back the emotional state in real time, allowing elderly people and foreign customers to receive appropriate services.

[1321] "Voice data" is a digital recording of a user's speech.

[1322] "Collection means" refers to the devices and techniques used to collect audio data.

[1323] "Preprocessing" refers to performing data processing such as noise removal and normalization on collected voice data.

[1324] A "speech recognition model" is an algorithm or machine learning model that analyzes voice data and converts it into text data.

[1325] "Text" is character data converted from audio data.

[1326] A "multilingual generative AI model" is an artificial intelligence model used to translate text input in one language into another language.

[1327] "Translation" is the conversion of text written in one language into another.

[1328] "Display means" refers to devices or techniques for visually presenting the translated text and emotional state to the user.

[1329] "Audio playback" refers to synthesizing text data into audio and providing it to the user audibly.

[1330] The "emotional state" indicates the emotional state recognized from the user's speech and facial expressions.

[1331] "Means of expert checking and supplementation" refers to a method in which an expert checks the accuracy of the translation results and provides corrections or additional information as necessary.

[1332] "Encryption" is a technology that converts data based on a specific algorithm to prevent unauthorized access by third parties.

[1333] "Cultural training" is the process by which generative AI models learn to adapt to multicultural environments.

[1334] This invention is a system for facilitating communication between elderly and foreign customers and store staff in brick-and-mortar stores. This system collects user speech data and performs preprocessing, translation, and emotion recognition using the following means:

[1335] 1. Audio data collection:

[1336] The server uses a terminal to collect voice data, such as a microphone built into smart glasses or a smartphone, which allows the server to capture customer speech in real time.

[1337] 2. Preprocessing the audio data:

[1338] The server then denoises and normalizes the received audio data using voice processing APIs from Amazon AWS and Google Cloud, making the data easier to analyze.

[1339] 3. Audio to text conversion:

[1340] The server then feeds the preprocessed audio data into a speech recognition model and converts it into text using an automatic speech recognition algorithm such as IBM Watson Speech to Text.

[1341] 4. Text translation:

[1342] The server translates the converted text using a multilingual generative AI model, using a translation engine such as the DeepL translation API, enabling smooth communication between different languages.

[1343] 5. Emotion recognition:

[1344] The server uses a natural language processing API (NLP) from Google Cloud to analyze the user's emotions from the text data. By identifying the user's emotional state, it can provide a more appropriate response.

[1345] 6. Displaying translation results and emotional state and playing audio:

[1346] The translated text and emotional state are displayed on the HUD (heads-up display) or smartphone screen, and the translation is played back via speech synthesis for real-time understanding by both the customer and store staff.

[1347] 7. Expert check and completion:

[1348] Experts check the accuracy of the translated content and provide corrections or additional information as needed, which increases the reliability of the system.

[1349] For example, when a store clerk asks, "Welcome. What can I do for you?", an elderly customer responds, "My feet have been hurting recently, so it's a bit difficult to walk." This speech data is collected, preprocessed, and converted into text using a speech recognition model. Next, this text is translated into English using the DeepL translation API, and the emotional state is identified as "negative" using Google Cloud's Natural Language API. Finally, the HUD displays, "My feet have been hurting recently, so it's a bit difficult to walk. Emotion score: -0.3," and the translation result is conveyed to the store clerk through audio playback.

[1350] An example of a prompt sentence for the generative AI model is as follows:

[1351] "Audio of a patient speaking Japanese in a hospital"

[1352] "output_language: English"

[1353] "output_emotion: emotion analysis"

[1354] "expected_output: Translation result: 'My feet have been hurting recently, so it's a bit difficult to walk.' Sentiment score: '-0.3'"

[1355] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1356] Step 1:

[1357] The user speaks into the microphone built into the smart glasses or smartphone. The device captures the customer's voice data and transmits it to the server in real time. The input is the customer's voice data, and the output is the digital voice data sent to the server.

[1358] Step 2:

[1359] The server denoises and normalises the received audio data. This process uses the audio processing APIs of Amazon AWS and Google Cloud. The input is the received audio data, and the output is the denoised and normalised audio data.

[1360] Step 3:

[1361] The server inputs the preprocessed audio data into a speech recognition model and converts it into text. IBM Watson Speech to Text is used here. The input is the preprocessed audio data, and the output is text data.

[1362] Step 4:

[1363] The server translates the converted text using a multilingual generative AI model. This process uses the DeepL translation API. The input is text data, and the output is the translated text.

[1364] Step 5:

[1365] The server uses a natural language processing API to analyze the user's emotional state from the translated text. Google Cloud's Natural Language API is used here. The input is the translated text data, and the output is the emotion score and emotion type.

[1366] Step 6:

[1367] The server displays the translated text and emotional state on a HUD (Heads-Up Display) or smartphone screen. The displayed content includes the translated text and the emotional score. The input is the translated text and the emotional state, and the output is the displayed information.

[1368] Step 7:

[1369] The server synthesizes the translation result and plays it back, allowing the user (store clerk) to listen to the translation in real time. The input is the translated text, and the output is audio data.

[1370] Step 8:

[1371] The expert (clerk) checks the displayed and played translation results and emotional state, and provides corrections or additional information as needed. The input is the translated text and emotional state, and the output is the corrected or supplemented information.

[1372] 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.

[1373] 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.

[1374] 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.

[1375] [Fourth embodiment]

[1376] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1377] 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.

[1378] 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).

[1379] 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.

[1380] 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.

[1381] 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).

[1382] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1383] 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.

[1384] 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.

[1385] 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.

[1386] 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.

[1387] 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.

[1388] 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."

[1389] This is a system to support smooth communication between elderly and foreign patients and medical professionals. The system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment.

[1390] Program implementation and processing explanation

[1391] Basic configuration

[1392] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[1393] Overview of the main measures that will emerge

[1394] 1. Collection of audio data

[1395] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[1396] 2. Preprocessing of audio data

[1397] Server: Performs noise reduction and normalization on the received audio data, removing unnecessary noise and converting the audio data into a format that is easier to analyze.

[1398] 3. Speech to text conversion

[1399] Server: The preprocessed speech data is analyzed using a speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[1400] 4. Text Translation

[1401] Server: The converted text data is translated in real time using a multilingual generative AI model. For example, the Japanese text "I have a headache" is translated into English as "I have a headache."

[1402] 5. Display and playback of translated text and audio

[1403] On your device: The translated text will be displayed on your device screen and played back using speech synthesis, if available. For example, the translation will be displayed on your device and the speech synthesis will say "I have a headache."

[1404] 6. Expert check and completion

[1405] User (medical professional): Checks the translation results and corrects or completes them as necessary. This improves the accuracy and reliability of the translation. For example, a medical professional corrects "windpipe" to "trachea."

[1406] 7. Data Encryption and Protection

[1407] Server: All data is protected with advanced encryption technology, for example, using the AES encryption algorithm, and data in transit is sent and received over secure protocols (TLS / SSL).

[1408] 8. Training a generative AI model

[1409] Server: Generative AI models are trained to understand cultural differences and sensitivities, for example, using datasets specific to Japanese customs and expressions.

[1410] Specific examples

[1411] Consider a scenario in which a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital.

[1412] 1. Voice input: The patient speaks to the terminal in English, "I have a headache."

[1413] 2. Data preprocessing and speech recognition: This speech data is sent to the server, where it is noise-reduced and converted to text using an ASR model.

[1414] 3. Text translation: A generative AI model translates the English phrase "I have a headache" into Japanese phrase "I have a headache."

[1415] 4. Display and audio output: The translated Japanese text is displayed on the doctor's device and played as audio if necessary.

[1416] 5. Expert review: A doctor reviews the translation and asks additional questions or makes a diagnosis.

[1417] The above steps will enable rapid and accurate medical communication that transcends language barriers.

[1418] The processing flow will be explained below.

[1419] Step 1:

[1420] Audio data collection

[1421] Device: The patient or healthcare professional speaks into the device. The device's built-in microphone captures the audio and begins recording.

[1422] Example of operation: When a patient says "I have a headache" to the device, the voice is recorded.

[1423] Step 2:

[1424] Sending audio data

[1425] Terminal: Streams recorded audio data to the server in real time.

[1426] How it works: Captured audio data is converted into digital form and sent over a network to a server.

[1427] Step 3:

[1428] Audio data preprocessing

[1429] Server: Performs noise reduction and normalization on the received audio data, converting it into a clear and consistent format.

[1430] How it works: A filtering algorithm is applied to remove background noise and even out volume variations.

[1431] Step 4:

[1432] Text conversion using speech recognition models

[1433] Server: The preprocessed speech data is input into the speech recognition model and converted into text.

[1434] Example of how it works: A speech recognition algorithm analyzes the characteristics of a voice and converts the speech "I have a headache" into the text "I have a headache."

[1435] Step 5:

[1436] Text translation

[1437] Server: Using a generative AI model, translate text data into the target language. Use a multilingual translation engine.

[1438] Example of how it works: The text "I have a headache" is translated into English as "I have a headache".

[1439] Step 6:

[1440] Display and playback of translation results

[1441] On Device: Displays translated text and audio on the user's device, optionally using speech synthesis to play the translation aloud.

[1442] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache" in English.

[1443] Step 7:

[1444] Checked and supplemented by experts

[1445] User (medical professional): Review the translated content and provide corrections or additional information as needed.

[1446] Example of how it works: A doctor checks the translation results and corrects any inaccuracies. For example, he or she asks an additional question in Japanese: "Please tell me the cause of my headache."

[1447] Step 8:

[1448] Data encryption and protection

[1449] Server: We encrypt all data sent and received and use advanced security protocols to prevent unauthorized access to your data by third parties.

[1450] How it works: The server encrypts data using the AES encryption algorithm, and the encrypted data is sent and received via the TLS protocol.

[1451] Step 9:

[1452] Training a generative AI model

[1453] Server: Trains AI models to understand cultural differences and multilingual environments, thereby reducing mistranslations and cultural misunderstandings.

[1454] How it works: An AI model trains on a server using a multilingual dataset to improve its ability to understand Japanese-specific nuances.

[1455] Example 1

[1456] 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."

[1457] In today's medical settings, communication barriers between elderly or foreign patients and medical professionals are a serious problem. It is particularly difficult to communicate smoothly with patients who have language differences, hearing impairments, or cultural backgrounds, which can delay appropriate diagnoses and treatment. A system that can resolve these issues and support accurate and prompt communication in medical settings is needed.

[1458] 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.

[1459] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for experts to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for noise reduction and normalization, means for inputting voice into a voice recognition model, and means for synthesizing voice, thereby enabling accurate and prompt communication between patients and medical professionals with different languages, hearing impairments, or cultural backgrounds.

[1460] An "audio data collection means" is a device or method for capturing and transmitting audio from a patient or medical professional to a server.

[1461] The "audio data preprocessing means" refers to a device or method for converting collected audio data into a form that is easy to analyze by performing processes such as noise reduction and normalization.

[1462] A "speech recognition model" is a machine learning algorithm that analyzes preprocessed speech data and converts it into text data.

[1463] A "multilingual generative AI model" is an artificial intelligence model used to translate between different languages ​​in real time.

[1464] "Display and playback means" refers to a device or method for displaying the translated text on the screen of the terminal and, if necessary, synthesizing it to speech and playing it back.

[1465] "Expert checking and completion means" refers to an interface and method for medical professionals to check the translation results and make corrections or completions as necessary.

[1466] A "data encryption means" is a device or method for encrypting data using an encryption algorithm to protect all data processed within the system.

[1467] A "means for training a generative AI model" is a device or method for training a generative AI model with a particular dataset so that the model understands cultural differences and sensitivities.

[1468] "Noise reduction and normalization means" refers to a device or method for reducing noise in audio data during pre-processing and adjusting the volume and pitch to a consistent level.

[1469] A "means for inputting speech into a speech recognition model" is a device or method for appropriately inputting preprocessed speech data into a speech recognition model.

[1470] "Speech synthesis means" refers to a device or method for generating speech based on text data and reproducing it.

[1471] This invention relates to a system that supports smooth communication between elderly and foreign patients and medical professionals. Specifically, it enables accurate and prompt medical treatment by providing voice, text, translation, and real-time medical information.

[1472] Fundamental System Configuration

[1473] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, and a terminal that displays and plays the converted and translated information. An interface is also provided for experts to check and complete the translation results. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[1474] Technical Details

[1475] Hardware and Basic Operation

[1476] The device is used by the patient or a medical professional and collects voice data through a built-in microphone. The device can be a smartphone, tablet, or dedicated voice collection device. The collected voice data is sent to a server in real time. For example, if a patient says, "I have a headache," the voice is captured instantly.

[1477] The server preprocesses the received audio data. The main processes are noise reduction and normalization. Noise reduction removes background noise, and normalization maintains the quality of the audio data.

[1478] The preprocessed speech data is then converted into text data using an Automatic Speech Recognition (ASR) model. For example, the Japanese phrase "I have a headache" is converted directly into text.

[1479] The converted text is then translated in real time by a multilingual generative AI model, pre-trained to understand cultural differences and sensitivities, for example, translating the Japanese text "Watashi wa tou ga kaiyuu" (I have a headache) into English.

[1480] The translated text and audio are then displayed and played back on the device. The translated text is displayed on the device screen and the translation is played back using speech synthesis. For example, the translated "I have a headache" is displayed and played back as audio.

[1481] The user (medical professional) can check the translation results through the interface provided by the system and make corrections or additions as necessary, thereby improving the accuracy and reliability of the translation. For example, a medical professional can correct "windpipe" to "trachea."

[1482] The server protects all data with advanced encryption technology, specifically using the AES encryption algorithm, and data in transit is sent and received via secure protocols (TLS / SSL).

[1483] Specific examples

[1484] Consider a case where a foreign patient is having difficulty communicating smoothly when undergoing treatment at a Japanese hospital. The following steps are taken:

[1485] 1. Voice input:

[1486] The patient speaks to the terminal in English, "I have a headache."

[1487] 2. Data preprocessing and speech recognition:

[1488] This audio data is sent to a server, where it is noise reduced and normalized before being converted to text using an ASR model.

[1489] 3. Text Translation:

[1490] The server uses a generative AI model to translate the English phrase "I have a headache" into Japanese phrase "I have a headache."

[1491] 4. Display and audio output:

[1492] The translated Japanese text is displayed on the doctor's device and the audio is played back as "I have a headache."

[1493] 5. Expert review:

[1494] The doctor will check the translated content and provide appropriate diagnosis and treatment.

[1495] Prompt Sentence Examples

[1496] Examples of prompts to enter into this system are:

[1497] "If a patient says, 'I have a headache,' in English, convert that audio data into Japanese text and speech for the doctor to review."

[1498] The above procedures and system configuration will enable rapid and accurate medical communication that transcends language barriers.

[1499] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1500] Step 1:

[1501] Audio data collection

[1502] The device captures the voice of the patient or medical professional. When the patient speaks into the device, the built-in microphone collects the voice and transmits it to the server in real time. The input is voice data, and the output is raw voice data sent to the server.

[1503] Specific behavior:

[1504] The patient speaks into the terminal's microphone, "I have a headache."

[1505] The device collects the voice data and sends it to the server.

[1506] Step 2:

[1507] Audio data preprocessing

[1508] The server performs noise reduction and normalization on the received audio data, converting it into a form that is easier to analyze. The input is raw audio data, and the output is preprocessed audio data.

[1509] Specific behavior:

[1510] The server receives the audio data.

[1511] Apply a noise reduction algorithm to remove background noise.

[1512] Normalizes the volume and pitch of audio data.

[1513] Step 3:

[1514] Speech-to-text conversion

[1515] The server inputs the preprocessed speech data into an Automatic Speech Recognition (ASR) model and converts it into text data. The input is the preprocessed speech data, and the output is text data.

[1516] Specific behavior:

[1517] The server inputs the speech data into the ASR model.

[1518] The ASR model analyzes the speech data and generates the text "I have a headache."

[1519] Step 4:

[1520] Text translation

[1521] The server translates the converted text using a generative AI model. The input is text data, and the output is translated text data.

[1522] Specific behavior:

[1523] The server inputs text into the generative AI model.

[1524] A generative AI model translates the Japanese text "I have a headache" into English "I have a headache."

[1525] Step 5:

[1526] View and play translated text and audio

[1527] The terminal displays the translated text received from the server on the screen and plays it back using speech synthesis if necessary. The input is the translated text data, and the output is the displayed text and speech data.

[1528] Specific behavior:

[1529] The terminal receives the translation result from the server.

[1530] The translated text is displayed on the screen.

[1531] Uses a speech synthesis engine to play the translated text aloud.

[1532] Step 6:

[1533] Checked and supplemented by experts

[1534] The user (medical professional) checks the translation results displayed on the terminal and makes corrections or completions as necessary. The input is the translated text data, and the output is the corrected or completed text data.

[1535] Specific behavior:

[1536] The medical professional checks the translation results displayed on the device's display.

[1537] If the translation is not accurate, correct it.

[1538] Step 7:

[1539] Data encryption and protection

[1540] The server encrypts all audio and text data. The input is the processed data and the output is the encrypted data.

[1541] Specific behavior:

[1542] The server encrypts the audio and text data using the AES encryption algorithm.

[1543] Encrypted data is sent and received using secure protocols (TLS / SSL).

[1544] Step 8:

[1545] Training a generative AI model

[1546] The server trains the generative AI model to understand cultural differences and sensitivities: the input is a specific culture and language dataset, and the output is an optimized generative AI model.

[1547] Specific behavior:

[1548] The server inputs specific culture and language datasets into the AI ​​model.

[1549] The server optimizes the parameters of the AI ​​model and performs training.

[1550] (Application example 1)

[1551] 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."

[1552] To ensure smooth communication between elderly and foreign patients and medical professionals, accurate real-time speech recognition and translation, as well as secure processing and transmission of the data, are necessary. However, current systems lack the ability to process voice data in real time, securely encrypt the data, and train generative AI models that take cultural differences into account, hindering accurate and prompt medical responses.

[1553] 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.

[1554] In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a speech recognition model, means for translating the converted text using a multilingual generative AI model, means for displaying and playing back the translated text, means for an expert to check and complete the accuracy of the translated text, means for encrypting data, means for training the generative AI model to take cultural differences into account, means for securely storing and transmitting the encrypted data, and means for processing the voice data collected by the terminal in real time, thereby enabling real-time and accurate medical communication between patients and medical professionals.

[1555] "Audio data" refers to data in which an audio signal is recorded as digital information.

[1556] A "collection means" is a device or system for acquiring and transmitting audio data to another processing unit.

[1557] The "preprocessing means" is a device or system that performs preprocessing such as noise removal and normalization on collected voice data.

[1558] A "voice recognition model" is an algorithm or AI model that analyzes voice data and converts it into a string of characters (text).

[1559] "Text" is digital data expressed as character information.

[1560] A "multilingual generative AI model" is an AI model that supports multiple languages ​​and translates from one language to another in real time.

[1561] A "translation tool" is a device or system that converts text in a particular language into another language.

[1562] "Display means" refers to a device or interface for visually showing the translation results to the user.

[1563] "Playback means" refers to a device or system that provides the translation result to the user as audio.

[1564] "Means for expert checking and correction" refers to a device or system that allows an expert to verify the accuracy of the translation results and correct them if necessary.

[1565] An "encryption means" is a device or system that encrypts data to protect the security of the data.

[1566] A "training means for a generative AI model to take cultural differences into account" is a device or system that trains a generative AI model to understand differences in cultural background and expression.

[1567] A "real-time processing means" is a device or system that instantly analyzes and processes audio data and provides results to the user in real time.

[1568] "Storage and transmission means" means a device or system that securely stores encrypted data and transmits it wherever needed.

[1569] To implement this invention, a system must be constructed by combining a series of hardware and software. The specific system configuration, processing procedures, and usage methods are described below.

[1570] System configuration

[1571] This system is a multi-function system for speech recognition, real-time translation, data encryption, and displaying and playing back the results.

[1572] Hardware used

[1573] 1. Collection device: A device for collecting voice data, such as a smartphone or head-mounted display. This device has a built-in highly sensitive microphone.

[1574] 2. Servers: High-performance servers are used for data pre-processing, speech recognition, text conversion, translation, and data encryption, including cloud servers.

[1575] 3. Display device: A device that displays the translated data and plays it back as audio. This includes smartphones and tablets.

[1576] Software used

[1577] 1. Speech recognition library: A library for converting voice data to text (e.g., speech_recognition).

[1578] 2. Translation library: A library for translating text into another language (e.g. googletrans).

[1579] 3. Cryptography library: A library for performing data encryption (e.g., the Fernet module for cryptography).

[1580] System processing explanation in natural language

[1581] The server receives voice data from a collection terminal, which is used by a patient or a medical professional to capture the voice data. The captured voice data undergoes pre-processing for noise reduction and normalization before being sent to the server.

[1582] The server converts the preprocessed speech data into text using a speech recognition library. This text data is translated into the specified language in real time using a multilingual generative AI model. For example, if a patient says "I have a headache" in English, this will be translated into Japanese as "I have a headache."

[1583] The translated text is sent to a display terminal where it is displayed on the screen and played back as audio. The system also includes an interface that allows experts to check the accuracy of the translation and make corrections or additions if necessary.

[1584] All data is encrypted using advanced encryption technology and transmitted over secure protocols. The server uses the AES encryption algorithm and TLS / SSL protocols to ensure data security.

[1585] Examples of concrete examples and prompts

[1586] Specific usage scenarios include the following:

[1587] 1. Example: When an English-speaking patient visits a Japanese hospital, they use their smartphone to communicate their symptoms to the doctor. When they speak into their smartphone, for example, "I have a headache," this voice data is translated into Japanese and displayed and played on the doctor's device.

[1588] 2. Example prompt:

[1589] I'm an English-speaking patient with a headache. I say "I have a headache" in English to my smartphone. Please translate the voice data into Japanese and display it so that my doctor can see it. Also, please encrypt the data to ensure its security.

[1590] This invention enables real-time and accurate medical communication between elderly and foreign patients and medical professionals.

[1591] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1592] Step 1:

[1593] The collection terminal captures the speech data uttered by the patient or medical professional. The input is the speech signal collected from the microphone of the collection terminal. This speech signal is converted into a digital format and subjected to noise reduction and normalization processing. The output is the pre-processed speech data.

[1594] Step 2:

[1595] The preprocessed voice data is sent from the collection terminal to the server. The input is the preprocessed voice data. The collection terminal sends the data to the server using a transmission protocol. The output is the voice data received at the server side.

[1596] Step 3:

[1597] The server performs speech recognition processing on the received audio data. The input is noise-reduced and normalized audio data. The server uses a speech recognition library (e.g., speech_recognition) to convert the audio data into text data. The output is text data.

[1598] Step 4:

[1599] The server translates the converted text data using a multilingual generative AI model. The input is text data generated by a speech recognition model. A translation library (e.g., GoogleTrans) is used to translate the text into the specified language in real time. The output is the translated text data.

[1600] Step 5:

[1601] The server encrypts the translated text data. The input is the translated text data. The data is encrypted using an encryption library (e.g. Fernet). The output is the encrypted text data.

[1602] Step 6:

[1603] Encrypted text data is sent from the server to the display terminal. The input is the encrypted text data. The server transfers the data using a secure protocol (e.g., TLS / SSL). The output is the encrypted text data received by the display terminal.

[1604] Step 7:

[1605] The display terminal decrypts encrypted text data received from the server. The input is the encrypted text data. The display terminal uses the encryption library to decrypt the data. The output is the decrypted text data.

[1606] Step 8:

[1607] The display terminal displays and plays back the decoded text data. The input is the decoded text data. The user interface displays the text data on the screen and uses speech synthesis to play the text aloud. The output is a translation result that can be recognized visually and audibly by the user.

[1608] Step 9:

[1609] The expert user uses a display terminal to check the accuracy of the translated text data and make corrections or completions as necessary. The input is the decoded and displayed text data. The user makes corrections through the interface, and the results are reflected in the system. The output is the corrected and completed accurate translation data.

[1610] This series of steps enables accurate and secure medical communication in real time.

[1611] 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.

[1612] This invention is a system for supporting smooth communication and emotion recognition between elderly and foreign patients and medical professionals. This system provides voice, text, translation, and real-time medical information, enabling accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it is possible to make diagnoses and treatments that take into account the patient's emotional state.

[1613] Program implementation and processing explanation

[1614] Basic configuration

[1615] The system consists of the following components: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, and a terminal that displays and plays back the converted and translated information. An interface is also provided for experts to check and complete the translation results and recognized emotions. The system meets high security standards and employs various security methods, including data encryption and confidentiality.

[1616] Overview of the main measures that will emerge

[1617] 1. Collection of audio data

[1618] Device: The patient or medical professional speaks into the device. The device's built-in microphone captures the voice and sends the data to the server. For example, a patient might say, "I have a headache."

[1619] 2. Sending audio data

[1620] Terminal: Streams recorded audio data to the server in real time. The captured audio data is converted into a digital format and sent over the network to the server.

[1621] 3. Preprocessing of audio data

[1622] Server: Noise reduction and normalization processes are performed on the received audio data, converting the audio into a clear and consistent format. Filtering algorithms are applied to remove background noise and even out volume variations.

[1623] 4. Emotion Recognition by Emotion Engine

[1624] Server: Analyzes the preprocessed audio and video data to recognize the user's emotional state. For example, it identifies emotions such as "anger," "sadness," and "surprise." The emotion recognition algorithm reads the user's emotions from their voice intonation and facial expressions.

[1625] 5. Speech to Text Conversion

[1626] Server: The preprocessed speech data is input into the speech recognition model and converted into text data. For example, the speech "I have a headache" is converted into text data.

[1627] 6. Text Translation

[1628] Server: Translates text data into the target language using a generative AI model. It uses a multilingual translation engine. For example, the text "I have a headache" is translated into English as "I have a headache."

[1629] 7. Display and playback of translation results and emotions

[1630] Terminal: The translated text and recognized emotion information are displayed on the user's terminal. If necessary, the translation result is played back aloud using speech synthesis. For example, the translation result is displayed on the doctor's terminal, and the speech synthesis reproduces "I have a headache" in English. The display screen also indicates the patient's emotional state (e.g., "anxiety").

[1631] 8. Expert check and completion

[1632] User (medical professional): Review the translated content and the perceived sentiment, and provide corrections or additional information as needed. This improves the accuracy and reliability of the translation. For example, a doctor may review the translation and ask a follow-up question such as "What is causing your headache?", as well as provide suggestions to ease the patient's anxiety.

[1633] 9. Data Encryption and Protection

[1634] Server: All data is protected with advanced encryption technology. Voice and text data is encrypted with the AES encryption algorithm, and data in transit is also securely transmitted via the TLS protocol.

[1635] 10. Training a generative AI model

[1636] Server: Trains AI models to understand cultural differences and multilingual environments. Using multilingual datasets, AI models learn and improve their ability to understand Japanese-specific nuances.

[1637] These steps not only enable fast and accurate medical communication across language barriers, but also provide a more humane response that takes into account the patient's emotional state.

[1638] The processing flow will be explained below.

[1639] Step 1:

[1640] Audio data collection

[1641] Device: The patient or healthcare professional speaks into the device, which has a built-in microphone that captures the voice and records the data in real time.

[1642] Example of operation: When a patient says "I have a headache" to the device, the voice is instantly recorded and saved as digital audio data.

[1643] Step 2:

[1644] Sending audio data

[1645] Terminal: Recorded audio data is sent to the server in streaming format.

[1646] How it works: Send captured audio data to a server using any Wi-Fi or 4G network within range.

[1647] Step 3:

[1648] Audio data preprocessing

[1649] Server: Performs noise reduction and normalization on the received audio data, making it easier to analyze.

[1650] How it works: Using a noise filtering algorithm on the server side to remove background noise and enhance the audio signal.

[1651] Step 4:

[1652] Emotion recognition by emotion engine

[1653] Server: Analyzes the pre-processed audio and video data (if necessary) and recognizes the user's emotions. An emotion recognition algorithm is used.

[1654] Example of how it works: The server analyzes the intonation, speed, and emphasized parts of the voice to identify emotions such as "anger," "sadness," and "surprise." It also uses facial expression data recognized using a camera.

[1655] Step 5:

[1656] Speech-to-text conversion

[1657] Server: The preprocessed speech data is input into the speech recognition model (ASR) and converted into corresponding text data.

[1658] Example of how it works: A speech recognition algorithm listens for speech characteristics and converts the speech "I have a headache" into text "I have a headache."

[1659] Step 6:

[1660] Text translation

[1661] Server: Using a generative AI model, the converted text data is translated into the required language using a multilingual translation engine.

[1662] Working example: Translate the Japanese phrase "I have a headache" into English.

[1663] Step 7:

[1664] Display and playback of translation results and emotional information

[1665] Device: The translated text and recognized emotion information are displayed on the user's device, and the translation result is played aloud using speech synthesis if necessary.

[1666] Example of operation: The translation result is displayed on the doctor's device, and the speech synthesis reproduces "I have a headache." Emotional information such as "Emotion: Anxiety" is also displayed on the display screen.

[1667] Step 8:

[1668] Checked and supplemented by experts

[1669] User (medical professional): Checks the content of the translated text and sentiment information, and makes corrections or asks additional questions as needed, thereby improving the accuracy of the diagnosis.

[1670] Example of how it works: A doctor checks the translation results and emotional information, and then asks an additional question in Japanese, such as "Please tell me the cause of my headache." The system translates this into English in real time.

[1671] Step 9:

[1672] Data encryption

[1673] Server: All data sent and received is protected with the AES encryption algorithm, and communication is carried out via the TLS / SSL protocol.

[1674] Example of operation: All data, including recorded voice data, translated text, and emotional information, is encrypted and transmitted over a secure communication channel.

[1675] Step 10:

[1676] Training a generative AI model

[1677] Server: Trains generative AI models to understand cultural differences and multilingual environments, using multilingual datasets for training.

[1678] How it works: The server trains an AI model using a dataset on Japanese culture to understand Japanese-specific expressions and nuances.

[1679] The above processing flow enables accurate and prompt communication between patients and medical professionals, and also enables responses that take into account the patient's emotional state.

[1680] Example 2

[1681] 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."

[1682] It is difficult for elderly and foreign patients to communicate smoothly with medical professionals, especially when it comes to accurately recognizing emotions and providing appropriate responses. Furthermore, in multilingual medical communication, it is necessary to ensure the accuracy and reliability of translations. Furthermore, the security of medical information is also a major issue.

[1683] 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.

[1684] In this invention, the server includes a means for transmitting voice data in real time, a means for performing noise reduction and normalization on the preprocessed voice data, and a means for analyzing the preprocessed voice data to recognize emotions. This enables medical communication that takes into account the user's emotions while overcoming language barriers. It also improves the accuracy and reliability of translations and ensures the security of medical information.

[1685] "Means for collecting voice data" refers to the function of capturing voices spoken by patients or medical professionals using the device's microphone.

[1686] The "means for transmitting audio data to a server in real time" refers to a mechanism for converting captured audio data into a digital format in real time and transmitting it to a server via a network.

[1687] The "noise reduction and normalization means" is a function that removes background noise from the received audio data and performs processing to equalize the volume.

[1688] "Means for recognizing emotions" refers to a technology that analyzes pre-processed voice data and identifies the user's emotional state from the intonation and tempo of the voice.

[1689] A "speech recognition model" is a technology that analyzes voice data and converts it into text format, often using automatic speech recognition (ASR) algorithms.

[1690] A "multilingual generative AI model" is an AI technology that supports multiple languages ​​and translates input text data into other languages.

[1691] "Display and playback means" refers to the functionality of providing the translated text and emotional information to the user visually and audibly.

[1692] "Means for expert checking and correction" refers to a system in which experts can check and correct the content of translated text and the results of emotion recognition.

[1693] "Means for encrypting data" refers to a method of protecting all data with advanced encryption technology, ensuring that data is sent and received securely during communication.

[1694] "Training measures to take cultural differences into account" refers to the learning process that generative AI models undergo to understand the cultural nuances of each language and translate or respond appropriately.

[1695] This invention is a system that enables smooth communication between elderly and foreign patients and medical professionals, and also enables emotion recognition. This system provides voice, text, translation, and real-time medical information to support accurate and prompt medical treatment. In addition, by combining it with an emotion engine that recognizes the user's emotions, it becomes possible to diagnose and treat patients taking into account their emotional state.

[1696] Basic configuration

[1697] The system consists of the following elements: a terminal that collects voice data, a server that preprocesses and analyzes the voice data, a server that recognizes the user's emotions using an emotion engine, a terminal that displays and plays back the converted and translated information, and an interface for experts to check and complete the translation results and recognized emotions. The system also employs data encryption and confidentiality technologies to meet high security standards.

[1698] Hardware and Software

[1699] 1. Device:

[1700] It is a device with a user interface and a built-in microphone for collecting and capturing voice data. The device also has a processor, memory, and a network module, and transmits the voice data to a server in real time.

[1701] 2. Server:

[1702] The server preprocesses the received voice data and performs emotion and speech recognition. It has a high-performance processor, memory, and data storage, and runs speech recognition models and generative AI models, which then perform noise reduction and data analysis.

[1703] 3. Software:

[1704] The speech recognition model running on the server uses an automatic speech recognition (ASR) algorithm to convert preprocessed speech data into text. The generative AI model is multilingual and translates text data into other languages. In particular, it employs generative AI technology that has a proven track record as a translation engine.

[1705] 4. Emotion Engine:

[1706] This engine identifies the user's emotions by analyzing the intonation and waveform of voice data, and recognizes a variety of emotional states (e.g., anger, sadness, anxiety).

[1707] 5. Security Technology:

[1708] The AES encryption algorithm is used to encrypt data, and data is sent and received securely via the TLS protocol during communication.

[1709] 6. Expert interface:

[1710] This interface allows experts to review the translation results and the perceived sentiment, and provide corrections or additional information as needed. Users can input appropriate actions, thereby improving the accuracy and reliability of the translation.

[1711] Specific examples

[1712] For a specific scenario, use the following prompt:

[1713] Prompt Sentence Examples

[1714] The patient is expected to say to the device in Japanese, "I have chest pain," and the doctor will understand this in English.

[1715] The device captures the patient's voice saying "I feel chest pain," converts it into digital format, and sends it to the server. The server performs noise reduction on the received voice data and then uses an emotion engine to recognize the emotion "anxiety." Next, a speech recognition model converts the voice into text, generating the text data "I feel chest pain." This text data is translated into English by a generative AI model, resulting in the result "I feel chest pain." The translation result and emotion information are displayed on the device and, if necessary, played back using a speech synthesis function. Finally, an expert reviews the result and makes corrections or asks additional questions as necessary. The expert can then ask an additional question, "When did the chest pain start?" and provide measures to alleviate the patient's anxiety.

[1716] As described above, this system not only enables fast and accurate medical communication across language barriers, but also provides a humane response that takes into account the patient's emotional state.

[1717] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1718] Step 1:

[1719] Audio data collection

[1720] Device: The patient or medical professional speaks into the device's microphone. Voice data is input in analog format, and the microphone captures the voice and converts it into digital format. Specifically, the device activates the microphone and records the patient saying, "I have a headache." The recorded voice data is stored in a buffer.

[1721] Input: Patient speech (e.g., "I have a headache")

[1722] Output: Digital audio data

[1723] Step 2:

[1724] Sending audio data

[1725] Terminal: Converts recorded audio data into digital format in real time and streams it to the server over the network. Specifically, the terminal packs audio data into packets and sends them to the server via the network module.

[1726] Input: Recorded digital audio data

[1727] Output: Audio data sent to the server

[1728] Step 3:

[1729] Audio data preprocessing

[1730] Server: Performs noise reduction and normalization on the received audio data to convert it into a clear, uniform format. Specifically, the server filters the audio data, removes background noise, and normalizes volume fluctuations.

[1731] Input: Audio data sent to the server

[1732] Output: Noise-reduced and normalized audio data

[1733] Step 4:

[1734] Emotion recognition by emotion engine

[1735] Server: Analyzes the preprocessed voice data and identifies the user's emotions from the intonation and waveform of the voice data. Specifically, the emotion engine analyzes the frequency components of the voice data and recognizes emotions such as "anger," "sadness," and "surprise."

[1736] Input: Preprocessed audio data

[1737] Output: Recognized emotion tag (e.g., "anxiety")

[1738] Step 5:

[1739] Speech-to-text conversion

[1740] Server: Using a speech recognition model, the preprocessed speech data is converted into text format. Specifically, the speech data is broken down into phonemes, which are then reconstructed into strings of characters using a language model to generate text data.

[1741] Input: Preprocessed audio data

[1742] Output: Text data (e.g., "I have a headache")

[1743] Step 6:

[1744] Text translation

[1745] Server: Translates text data into other languages ​​using a generative AI model. Specifically, the generative AI model first analyzes the input text and then generates a translation based on the grammar and vocabulary of the target language. For example, "I have a headache" is translated into "I have a headache."

[1746] Input: Text data (e.g., "I have a headache")

[1747] Output: Translated text data (e.g., "I have a headache")

[1748] Step 7:

[1749] Display and playback of translation results and emotions

[1750] Terminal: The translated text and emotion information are displayed on the user's terminal and played back using a speech synthesis function as needed. Specifically, the terminal displays the text on the screen and uses a speech synthesis engine to play back the translation result aloud. For example, the translation result is displayed on the doctor's terminal and the phrase "I have a headache" is played back.

[1751] Input: Translated text data and recognized sentiment tags

[1752] Output: Text displayed on the screen, translation results played back aloud, and emotional information displayed

[1753] Step 8:

[1754] Checked and supplemented by experts

[1755] User (medical professional): Checks the translated content and the recognized sentiment, and provides corrections or additional information as needed. Specifically, the professional reviews the displayed translation, makes corrections if necessary, and enters a follow-up question. For example, a doctor might enter a follow-up question such as, "What causes headaches?"

[1756] Input: Translated text data and recognized sentiment tags

[1757] Output: Corrected translation and additional questions

[1758] Step 9:

[1759] Data encryption and protection

[1760] Server: All data is protected with advanced encryption technology. Specifically, voice and text data is encrypted with the AES encryption algorithm, and data in transit is securely transmitted and received using the TLS protocol.

[1761] Input: Audio data, text data, emotion data

[1762] Output: AES encrypted data, secure communication via TLS protocol

[1763] Step 10:

[1764] Training a generative AI model

[1765] Server: Trains the AI ​​model to understand cultural differences and multilingual environments. Specifically, the AI ​​model learns using multilingual datasets to improve its ability to understand Japanese nuances and cultural background.

[1766] Input: Multilingual dataset

[1767] Output: Generative AI models that are culturally and multilingually responsive

[1768] (Application example 2)

[1769] 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."

[1770] Elderly and foreign customers can sometimes find it difficult to communicate smoothly with store staff. In such situations, customers may not receive appropriate service due to language barriers or a lack of understanding of their emotions. Particularly in physical stores, where appropriate responses are required in real time, a system that solves these issues is necessary.

[1771] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data, means for preprocessing the collected voice data, means for converting the preprocessed voice data into text using a voice recognition model, means for translating the converted text using a multilingual generation AI model, means for displaying the translated text and emotional state, and means for audibly playing back the translated text and emotional state. This makes it possible to translate voice data into multiple languages ​​and display and play back the emotional state in real time, allowing elderly people and foreign customers to receive appropriate services.

[1772] "Voice data" is a digital recording of a user's speech.

[1773] "Collection means" refers to the devices and techniques used to collect audio data.

[1774] "Preprocessing" refers to performing data processing such as noise removal and normalization on collected voice data.

[1775] A "speech recognition model" is an algorithm or machine learning model that analyzes voice data and converts it into text data.

[1776] "Text" is character data converted from audio data.

[1777] A "multilingual generative AI model" is an artificial intelligence model used to translate text input in one language into another language.

[1778] "Translation" is the conversion of text written in one language into another.

[1779] "Display means" refers to devices or techniques for visually presenting the translated text and emotional state to the user.

[1780] "Audio playback" refers to synthesizing text data into audio and providing it to the user audibly.

[1781] The "emotional state" indicates the emotional state recognized from the user's speech and facial expressions.

[1782] "Means of expert checking and supplementation" refers to a method in which an expert checks the accuracy of the translation results and provides corrections or additional information as necessary.

[1783] "Encryption" is a technology that converts data based on a specific algorithm to prevent unauthorized access by third parties.

[1784] "Cultural training" is the process by which generative AI models learn to adapt to multicultural environments.

[1785] This invention is a system for facilitating communication between elderly and foreign customers and store staff in brick-and-mortar stores. This system collects user speech data and performs preprocessing, translation, and emotion recognition using the following means:

[1786] 1. Audio data collection:

[1787] The server uses a terminal to collect voice data, such as a microphone built into smart glasses or a smartphone, which allows the server to capture customer speech in real time.

[1788] 2. Preprocessing the audio data:

[1789] The server then denoises and normalizes the received audio data using voice processing APIs from Amazon AWS and Google Cloud, making the data easier to analyze.

[1790] 3. Audio to text conversion:

[1791] The server then feeds the preprocessed audio data into a speech recognition model and converts it into text using an automatic speech recognition algorithm such as IBM Watson Speech to Text.

[1792] 4. Text translation:

[1793] The server translates the converted text using a multilingual generative AI model, using a translation engine such as the DeepL translation API, enabling smooth communication between different languages.

[1794] 5. Emotion recognition:

[1795] The server uses a natural language processing API (NLP) from Google Cloud to analyze the user's emotions from the text data. By identifying the user's emotional state, it can provide a more appropriate response.

[1796] 6. Displaying translation results and emotional state and playing audio:

[1797] The translated text and emotional state are displayed on the HUD (heads-up display) or smartphone screen, and the translation is played back via speech synthesis for real-time understanding by both the customer and store staff.

[1798] 7. Expert check and completion:

[1799] Experts check the accuracy of the translated content and provide corrections or additional information as needed, which increases the reliability of the system.

[1800] For example, when a store clerk asks, "Welcome. What can I do for you?", an elderly customer responds, "My feet have been hurting recently, so it's a bit difficult to walk." This speech data is collected, preprocessed, and converted into text using a speech recognition model. Next, this text is translated into English using the DeepL translation API, and the emotional state is identified as "negative" using Google Cloud's Natural Language API. Finally, the HUD displays, "My feet have been hurting recently, so it's a bit difficult to walk. Emotion score: -0.3," and the translation result is conveyed to the store clerk through audio playback.

[1801] An example of a prompt sentence for the generative AI model is as follows:

[1802] "Audio of a patient speaking Japanese in a hospital"

[1803] "output_language: English"

[1804] "output_emotion: emotion analysis"

[1805] "expected_output: Translation result: 'My feet have been hurting recently, so it's a bit difficult to walk.' Sentiment score: '-0.3'"

[1806] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1807] Step 1:

[1808] The user speaks into the microphone built into the smart glasses or smartphone. The device captures the customer's voice data and transmits it to the server in real time. The input is the customer's voice data, and the output is the digital voice data sent to the server.

[1809] Step 2:

[1810] The server denoises and normalises the received audio data. This process uses the audio processing APIs of Amazon AWS and Google Cloud. The input is the received audio data, and the output is the denoised and normalised audio data.

[1811] Step 3:

[1812] The server inputs the preprocessed audio data into a speech recognition model and converts it into text. IBM Watson Speech to Text is used here. The input is the preprocessed audio data, and the output is text data.

[1813] Step 4:

[1814] The server translates the converted text using a multilingual generative AI model. This process uses the DeepL translation API. The input is text data, and the output is the translated text.

[1815] Step 5:

[1816] The server uses a natural language processing API to analyze the user's emotional state from the translated text. Google Cloud's Natural Language API is used here. The input is the translated text data, and the output is the emotion score and emotion type.

[1817] Step 6:

[1818] The server displays the translated text and emotional state on a HUD (Heads-Up Display) or smartphone screen. The displayed content includes the translated text and the emotional score. The input is the translated text and the emotional state, and the output is the displayed information.

[1819] Step 7:

[1820] The server synthesizes the translation result and plays it back, allowing the user (store clerk) to listen to the translation in real time. The input is the translated text, and the output is audio data.

[1821] Step 8:

[1822] The expert (clerk) checks the displayed and played translation results and emotional state, and provides corrections or additional information as needed. The input is the translated text and emotional state, and the output is the corrected or supplemented information.

[1823] 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.

[1824] 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.

[1825] 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.

[1826] 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.

[1827] 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.

[1828] 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.

[1829] 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).

[1830] 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.

[1831] 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."

[1832] 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.

[1833] 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).

[1834] 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.

[1835] 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.

[1836] 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.

[1837] 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.

[1838] 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.

[1839] 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.

[1840] 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.

[1841] 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.

[1842] 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.

[1843] 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.

[1844] The following is further disclosed regarding the above embodiment.

[1845] (Claim 1)

[1846] means for collecting audio data;

[1847] means for pre-processing the collected audio data;

[1848] means for converting the preprocessed speech data into text using a speech recognition model;

[1849] A means for translating the converted text using a multilingual generative AI model; and

[1850] a means for displaying and playing the translated text;

[1851] A means for expert checking and completion of the accuracy of the translated text;

[1852] means for encrypting the data;

[1853] means for training the generative AI model to take into account cultural differences;

[1854] A system including:

[1855] (Claim 2)

[1856] 2. The system according to claim 1, wherein the means for collecting voice data captures voice data from a terminal of a patient or a medical professional and transmits the voice data to a server.

[1857] (Claim 3)

[1858] 10. The system of claim 1, wherein the speech recognition model converts speech data to text using an automatic speech recognition (ASR) algorithm.

[1859] "Example 1"

[1860] (Claim 1)

[1861] means for collecting audio data;

[1862] means for pre-processing the collected audio data;

[1863] means for converting the preprocessed speech data into text using a speech recognition model;

[1864] A means for translating the converted text using a multilingual generative AI model; and

[1865] a means for displaying and playing the translated text;

[1866] A means for expert checking and completion of the accuracy of the translated text;

[1867] means for encrypting the data;

[1868] means for training the generative AI model to take into account cultural differences;

[1869] a means for noise reduction and normalization;

[1870] means for inputting speech into a speech recognition model;

[1871] means for performing speech synthesis;

[1872] A system including:

[1873] (Claim 2)

[1874] 2. The system according to claim 1, wherein the means for collecting voice data captures voice data from a terminal of a patient or a medical professional and transmits the voice data to a server.

[1875] (Claim 3)

[1876] 10. The system of claim 1, wherein the speech recognition model converts speech data to text using an automatic speech recognition (ASR) algorithm.

[1877] "Application Example 1"

[1878] (Claim 1)

[1879] means for collecting audio data;

[1880] means for pre-processing the collected audio data;

[1881] means for converting the preprocessed speech data into text using a speech recognition model;

[1882] A means for translating the converted text using a multilingual generative AI model; and

[1883] a means for displaying and playing the translated text;

[1884] A means for expert checking and completion of the accuracy of the translated text;

[1885] means for encrypting the data;

[1886] means for training the generative AI model to take into account cultural differences;

[1887] A means of securely storing and transmitting encrypted data;

[1888] means for processing the collected voice data in real time at the terminal;

[1889] A system including:

[1890] (Claim 2)

[1891] 2. The system according to claim 1, wherein the means for collecting voice data captures voice data from a terminal of a patient or a medical professional and transmits the voice data to a server, which processes the voice data in real time.

[1892] (Claim 3)

[1893] 2. The system of claim 1, wherein the speech recognition model converts speech data into text using an automatic speech recognition (ASR) algorithm and performs multilingual translation using a generative AI model.

[1894] "Example 2: Combining Emotion Engines"

[1895] (Claim 1)

[1896] means for collecting audio data;

[1897] means for transmitting the collected voice data to a server in real time;

[1898] means for performing noise reduction and normalization on the preprocessed audio data;

[1899] means for analyzing preprocessed speech data to recognize emotions;

[1900] means for converting the preprocessed speech data into text using a speech recognition model;

[1901] A means for translating the converted text using a multilingual generative AI model; and

[1902] means for displaying and playing back the translated text and emotion information;

[1903] A means for expert checking and completion of the accuracy of the translated text;

[1904] means for encrypting the data;

[1905] A means to train generative AI models to take cultural differences into account; and

[1906] A system including:

[1907] (Claim 2)

[1908] 2. The system according to claim 1, wherein the means for collecting voice data captures voice data from a terminal of a patient or a medical professional and transmits the voice data to a server in real time.

[1909] (Claim 3)

[1910] 2. The system of claim 1, wherein the speech recognition model converts speech data into text using an automatic speech recognition (ASR) algorithm, and the emotion recognition means analyzes intonation and waveform of the speech data to identify emotions.

[1911] "Application example 2 when combining emotion engines"

[1912] (Claim 1)

[1913] means for collecting audio data;

[1914] means for pre-processing the collected audio data;

[1915] means for converting the preprocessed speech data into text using a speech recognition model;

[1916] A means for translating the converted text using a multilingual generative AI model; and

[1917] a means for displaying and playing the translated text;

[1918] A means for expert checking and completion of the accuracy of the translated text;

[1919] means for encrypting the data;

[1920] means for training the generative AI model to take into account cultural differences;

[1921] a means for displaying the translated text and emotional state;

[1922] means for audibly reproducing the translated text and emotional state;

[1923] A system including:

[1924] (Claim 2)

[1925] 2. The system according to claim 1, wherein the voice data collection means captures voice data from a user's terminal and transmits the voice data to a server.

[1926] (Claim 3)

[1927] 10. The system of claim 1, wherein the speech recognition model converts speech data to text using an automatic speech recognition algorithm. [Explanation of symbols]

[1928] 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 collecting audio data; means for pre-processing the collected audio data; means for converting the preprocessed speech data into text using a speech recognition model; A means for translating the converted text using a multilingual generative AI model; and a means for displaying and playing the translated text; A means for expert checking and completion of the accuracy of the translated text; means for encrypting the data; means for training the generative AI model to take into account cultural differences; A system including:

2. 2. The system according to claim 1, wherein the means for collecting voice data captures voice data from a terminal of a patient or a medical professional and transmits the voice data to a server.

3. 10. The system of claim 1, wherein the speech recognition model converts speech data to text using an automatic speech recognition (ASR) algorithm.

Images