System

Abstract

A system is provided.SOLUTION: A system comprising: means for receiving a user's voice stream and converting it into neutral voice; means for obtaining a user's video stream and sequentially processing it to detect a user's face; means for superimposing a neutral avatar on the detected face; and means for outputting the processed voice stream and video stream.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] The existence of gender bias in modern online interviews causes many problems. This bias is caused by interviewers unconsciously evaluating candidates' abilities and aptitudes based on their gender. To solve this problem, effective and practical methods to eliminate gender bias are needed. However, consciously eliminating bias is not easy, and artificial efforts alone are often insufficient. [Means for solving the problem]

[0005] The present invention provides a system that includes a means for receiving a user's voice stream and converting it to a gender-neutral voice, a means for acquiring the user's video stream and sequentially processing it to detect the user's face, a means for superimposing a gender-neutral avatar on the detected face, and a means for outputting the processed voice stream and video stream, thereby realizing an environment in which interviewers can evaluate candidates without gender bias.

[0006] This allows for fair and objective evaluation of online interviews, regardless of actual gender. Specifically, the system applies an algorithm to convert the audio stream into a gender-neutral voice, a face detection algorithm to detect faces, and a resized avatar image to overlay on the face, enhancing the system's feasibility and effectiveness.

[0007] An "audio stream" is a data stream for continuously transmitting or receiving a user's spoken voice as digital data.

[0008] A "gender-neutral voice" is a voice with vocal characteristics that are not biased toward a specific gender, and has been adjusted for pitch and tone.

[0009] A "video stream" is a data flow for continuously transmitting or receiving a user's video as digital data.

[0010] "Face detection" is the process of identifying and locating a user's face in frames within a video stream.

[0011] An "avatar" is a virtual facial image with a neutral appearance that is displayed in place of the user's face.

[0012] "Superimposition" is the process of superimposing another image (in this case, an avatar) on top of the detected face.

[0013] An "algorithm" is a set of rules or steps that defines a procedure or computational process for solving a particular problem.

[0014] "Sequential processing" refers to the method of analyzing and processing each frame in a video stream in order.

[0015] "Resizing" is a process of changing the size of an image. [Brief explanation of the drawings]

[0016] [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

[0017] 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.

[0018] First, the terms used in the following description will be explained.

[0019] 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).

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 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.

[0027] 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).

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 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.

[0034] 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.

[0035] 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.

[0036] 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."

[0037] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[0038] server

[0039] The server has the function of receiving the voice stream sent by the user. The received voice stream is converted into a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the input voice to generate a gender-neutral voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's terminal.

[0040] For example, if a user speaks in a high-pitched voice, the voice will be changed by the server to an appropriate pitch and transmitted to the interviewer in a neutral tone.

[0041] Terminal

[0042] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[0043] Specifically, a face detection algorithm identifies the position and size of the face within the frame, then resizes the avatar image to fit that size. The resized avatar is then overlaid on top of the original face image. This modified frame is displayed to both the user and the interviewer.

[0044] For example, if the user is smiling in the displayed frame, the avatar will be adjusted to appear smiling as well.

[0045] User

[0046] Users log in to the online interview system and grant permission to use the camera and microphone, which allows the system to capture their voice and video. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed on the video. This allows users to be interviewed without any bias regarding their gender.

[0047] For example, when a user introduces themselves in an online interview, their voice will be transmitted as a gender-neutral voice and a gender-neutral avatar will be displayed on the video screen. The interviewer will see and hear this gender-neutral voice and avatar, allowing the interview to proceed without being influenced by gender bias.

[0048] This invention allows interviewers to evaluate candidates without gender bias, thereby realizing fair and objective interviews.

[0049] The processing flow will be explained below.

[0050] Processing on the server

[0051] Step 1:

[0052] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[0053] Step 2:

[0054] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[0055] Step 3:

[0056] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[0057] Processing on the device

[0058] Step 1:

[0059] The device receives a video stream from the user's camera and uses OpenCV and other libraries to capture video frames from the camera.

[0060] Step 2:

[0061] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[0062] Step 3:

[0063] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[0064] Step 4:

[0065] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[0066] User processing

[0067] Step 1:

[0068] The user logs into the online interview system, accessing the system using a designated application or web portal.

[0069] Step 2:

[0070] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[0071] Step 3:

[0072] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[0073] The above processing steps enable online interviews that are free of gender bias.

[0074] Example 1

[0075] 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."

[0076] In conventional online interview systems, gender bias can affect evaluations. A particular problem is that interviewers may unconsciously have preconceived notions based on the candidate's gender. This makes it difficult to conduct fair and objective evaluations, and it is difficult to accurately determine the candidate's true abilities and aptitude.

[0077] 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.

[0078] In this invention, the server includes means for receiving the user's voice data and converting it into a gender-neutral voice, means for acquiring the user's video data and sequentially processing it to detect the user's face, and means for superimposing a gender-neutral virtual character on the detected face, thereby eliminating gender bias and enabling fair and objective interviews.

[0079] A "user" is a person who uses the online interview system, accesses the system, and provides their own audio and video.

[0080] "Audio Data" refers to audio information collected through the user's microphone, which is sent to the server for processing.

[0081] "Gender-neutral voice" refers to voices with pitch and tone adjusted so that they are not associated with a specific gender.

[0082] "Video data" refers to video information collected through a user's camera, material that is processed frame by frame.

[0083] "Sequential processing" refers to a method of processing input data sequentially in real time.

[0084] "Face detection" refers to the process of identifying the location and size of a user's face from frames within video data.

[0085] "Virtual characters" are neutral images or animations that are superimposed on the user's face and are designed to avoid any gender-related associations.

[0086] "Superimposition" is the process of superimposing one image or object on top of another.

[0087] "Gender bias" is a term that refers to preconceived notions or prejudices against a particular gender, and is an unfair factor that influences evaluations and judgments.

[0088] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[0089] server

[0090] The server has hardware and software to receive the voice data sent by the user. After receiving the voice data, the server uses the Vocoder algorithm to adjust the pitch and tone of the voice and convert it into a neutral voice. This converted voice data is then sent to the interviewer's device in real time.

[0091] For example, if a user provides a normal speaking voice, the voice is analyzed by the server and normalized to a standard pitch and tone.

[0092] Terminal

[0093] The device is equipped with hardware and software to capture video data from the user's camera. The captured video data is then processed sequentially using a face detection algorithm, such as OpenCV or Dlib, to identify the location and size of the user's face and overlay a gender-neutral virtual character on it.

[0094] This superimposed video data is displayed in real time to both the user and the interviewer. For example, when the user speaks with a smile, the virtual character is adjusted so that it also appears smiling based on that facial expression.

[0095] User

[0096] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured and processed on the server and device. The user's voice is converted to a gender-neutral voice, and a gender-neutral virtual character is superimposed on the video. This system allows the user to be interviewed without any gender bias.

[0097] For example, when a user introduces themselves in an online interview, their voice is converted to a gender-neutral voice on the server side, and a gender-neutral virtual character is displayed on the video screen. Through this gender-neutral voice and virtual character, the interviewer can conduct the interview without being influenced by gender bias.

[0098] Examples of prompt statements

[0099] Explain an online interview system to eliminate gender bias. Describe in detail the specific operation and examples of each component (server, terminal, user).

[0100] This invention is expected to enable fair and objective interviews and improve the accuracy of evaluations.

[0101] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0102] Step 1:

[0103] The user logs in to the online interview system. They enter their user ID and password on the login screen to access the system. The system then authenticates the user and sets up the system to start the interview session. The input is the user's authentication information, and the output is the system access rights after successful login.

[0104] Step 2:

[0105] The user grants permission to access the camera and microphone. The system displays a pop-up message, and the user must click "Allow" to access the device. The input is the user's permission, and the output is the connection status of the camera and microphone.

[0106] Step 3:

[0107] The user's device acquires video data from the camera and audio data from the microphone. These data are sent to the server as input data. For example, the camera captures video at 30 frames per second, and the microphone captures audio in real time. The input is the video and audio data, and the output is the data sent to the server.

[0108] Step 4:

[0109] The server receives the transmitted voice data and converts it into a gender-neutral voice using the Vocoder algorithm, adjusting the pitch and tone of the voice in real time to eliminate gender cues. The input is the original voice data, and the output is the converted gender-neutral voice data.

[0110] Step 5:

[0111] The converted neutral voice data is sent from the server to the interviewer's device in real time. The input is neutral voice data, and the output is the data sent to the interviewer's device. For example, if the user speaks in a high-pitched voice, the voice will be converted to a flat, neutral voice and delivered to the interviewer.

[0112] Step 6:

[0113] The device processes the video data acquired from the camera in real time and applies a face detection algorithm (e.g., OpenCV or Dlib) to identify the position and size of the user's face. The input is the original video data, and the output is the position and size information of the face.

[0114] Step 7:

[0115] Based on the detected face position and size, an image processing algorithm generates a gender-neutral virtual character and superimposes it on the original face image. The input is face position and size information, and the output is video data with the virtual character superimposed. For example, if a user is smiling while talking, the gender-neutral virtual character will also be displayed smiling.

[0116] Step 8:

[0117] The processed video data is displayed in real time to both the user and the interviewer. The input is video data with a virtual character superimposed, and the output is a video feed displayed on the user's and interviewer's devices.

[0118] Step 9:

[0119] During the interview, the interviewer asks the user questions through a neutral voice and virtual character, and the user answers. This repeated process is carried out in real time on the server and the terminal. The input is the interviewer's questions and the user's answers, and the output is real-time alternating communication between them.

[0120] Step 10:

[0121] After the interview is over, the user presses the "End" button to end the interview. This causes the server to stop processing the audio and video streams and end the session. The input is the end command, and the output is the stopped state of the audio and video streams.

[0122] (Application example 1)

[0123] 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."

[0124] Audio and video neutralization technology was invented to eliminate gender bias in online interview systems. Similarly, various voice operations and interactions are increasingly required in autonomous vehicles, where gender bias also presents a problem. Conventional in-vehicle interactive systems can contain unconscious gender bias in voice assistants and display displays, resulting in a lack of fairness for passengers and drivers.

[0125] 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.

[0126] In this invention, the server includes means for receiving a user's voice stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for outputting the processed voice stream and video stream to an interactive system of the autonomous vehicle, means for outputting the user's gender-neutral voice to an in-vehicle speaker, and means for displaying the gender-neutral avatar on an in-vehicle display using face detection, thereby eliminating gender bias in voice operations and display views in the autonomous vehicle and enabling fair and neutral interactions.

[0127] "User" refers to a person who uses a system or application.

[0128] An "audio stream" refers to a continuously transmitted stream of audio data.

[0129] "Gender-neutral voice" refers to a voice that has vocal characteristics that are not biased towards a specific gender.

[0130] A "video stream" refers to a continuously transmitted stream of video data.

[0131] "Face detection" refers to the process of recognizing the location and size of a face in a particular frame within a video stream.

[0132] A "gender-neutral avatar" refers to a virtual character whose visual elements are not gender-biased.

[0133] An "autonomous vehicle" refers to a vehicle that drives autonomously without the need for driver intervention.

[0134] An "interactive system" refers to a system that allows two-way communication with the user.

[0135] "In-vehicle speaker" refers to an audio output device installed inside a vehicle.

[0136] "Server" refers to a computer system that provides resources and processes data over a network.

[0137] "Sequential processing" refers to a method of processing data one by one in succession.

[0138] An "algorithm" refers to a method of solving a problem using specific computational procedures.

[0139] "Display" refers to a device or screen for displaying images or videos.

[0140] The present invention provides a technology for eliminating gender bias in interactive systems within autonomous vehicles. Specific embodiments of this system are described below.

[0141] Hardware and software used

[0142] 1. Hardware

[0143] Server: Located remotely and responsible for processing the audio and video streams.

[0144] Vehicle microphone: A device for capturing the user's voice stream.

[0145] Dash camera: A device for capturing a user's video stream.

[0146] Car speaker: A device for outputting converted neutral audio.

[0147] In-car display: A screen for displaying a gender-neutral avatar in the car.

[0148] 2. Software

[0149] Speech recognition engine (e.g., Google® Speech-to-Text API, Amazon Transcribe): Recognizes the user's voice stream and converts it into text.

[0150] Voice conversion algorithms (e.g., Voice Conversion NN(VCNN)): convert a voice stream into a neutral voice.

[0151] Face detection libraries (e.g. OpenCV): Detect faces in the captured video stream.

[0152] Avatar generation engine (e.g. Unity): Generates gender-neutral avatars based on face detection results.

[0153] User interface frameworks (e.g., React Native, Flutter®): for building the front end of in-car interactive systems.

[0154] Server Procedure

[0155] 1. Audio capture and conversion

[0156] The server receives the audio stream transmitted from the vehicle microphone.

[0157] The received audio stream is converted to a gender-neutral voice using a voice conversion algorithm, which adjusts pitch and tone to eliminate gender bias.

[0158] The converted neutral voice is output to the car's speakers.

[0159] 2. Video Stream Processing

[0160] The server receives the video stream transmitted from the in-car camera.

[0161] Each frame in the video stream is processed in real time using a face detection library.

[0162] A gender-neutral avatar generated by an avatar generation engine is superimposed on the detected face.

[0163] The generated avatar is displayed on the in-car display.

[0164] Specific examples

[0165] Audio capture and conversion examples

[0166] When a user says "Start navigation" to the in-car assistant, the server converts this speech into a gender-neutral voice in real time. The in-car speaker plays "Start navigation" in a consistent, gender-neutral tone. This real-time conversion is achieved using a voice conversion algorithm.

[0167] Video stream processing example

[0168] When a user enters a vehicle, the dashcam starts capturing the video stream. The server detects faces in each frame and displays a neutral avatar. For example, if the user is smiling, the avatar will adjust to smile as well.

[0169] Example prompts

[0170] "Convert to gender-neutral voice. This voice is a command for your in-car assistant."

[0171] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0172] Step 1: Capture the audio stream

[0173] The user speaks a voice command to the in-car assistant.

[0174] Input: User's voice captured by the vehicle's microphone.

[0175] Processing: The vehicle microphone sends the audio stream to the server.

[0176] Output: Real-time audio stream sent to the server.

[0177] Step 2: Neutralizing the Audio Stream

[0178] The server inputs the received audio stream into a voice conversion algorithm.

[0179] Input: The audio stream sent to the server.

[0180] Processing: A voice conversion algorithm (e.g., Voice Conversion NN (VCNN)) adjusts the pitch and tone of the voice to convert it into a neutral voice.

[0181] Output: Neutral converted audio stream.

[0182] Step 3: Output audio

[0183] The server sends a neutral audio stream to the car's speakers.

[0184] Input: Neutral converted audio stream.

[0185] Processing: A neutral audio stream is sent to the car speakers.

[0186] Output: Neutral voice played through the car speakers.

[0187] Step 4: Capturing the video stream

[0188] While the user is inside the car, the on-board camera captures a video stream.

[0189] Input: Video stream captured by the dashcam.

[0190] Processing: The video stream is sent to the server.

[0191] Output: Real-time video stream sent to the server.

[0192] Step 5: Face detection

[0193] The server inputs the received video stream into a face detection algorithm.

[0194] Input: The video stream sent to the server.

[0195] Processing: A face detection algorithm (e.g. OpenCV) detects the position and size of the user's face in each frame.

[0196] Output: Face position and size information for each frame.

[0197] Step 6: Avatar generation

[0198] The server inputs the face detection information into an avatar generation engine to generate a gender-neutral avatar.

[0199] Input: Face position and size information.

[0200] Processing: An avatar generation engine (e.g., Unity) generates a gender-neutral avatar based on the detected facial features and overlays it on each frame.

[0201] Output: A frame with an androgynous avatar superimposed.

[0202] Step 7: Output the video stream

[0203] The server transmits the video stream with the avatar superimposed on it to an in-car display.

[0204] Input: A video frame with an androgynous avatar superimposed on it.

[0205] Processing: The video stream is sent to the in-car display.

[0206] Output: A video stream with a gender-neutral avatar displayed on the in-car display.

[0207] 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.

[0208] The present invention relates to an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[0209] Processing on the server

[0210] The server receives the voice stream sent from the user's device. The received voice stream is converted to a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the voice to create a gender-neutral voice. The converted voice stream is then sent to the interviewer's device.

[0211] Furthermore, the server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, and this recognized emotion data is also sent to the interviewer's device.

[0212] For example, if a user speaks in an excited tone, the emotion engine will recognize this as "excited" and convey that information to the interviewer.

[0213] Processing on the device

[0214] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[0215] Furthermore, the device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, and changes the avatar's facial expression based on the recognized emotions.

[0216] For example, if the user is smiling, the emotion engine will recognize this as "happiness" and adjust the avatar to smile as well.

[0217] User processing

[0218] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the system to capture the user's voice and video. The user's voice is converted to a neutral voice and a neutral avatar is displayed on the video. Furthermore, the system recognizes the user's emotions and relays this information to the interviewer in real time.

[0219] For example, when a user introduces themselves in an online interview, their voice is conveyed as a neutral voice and a neutral avatar is displayed on the video screen. If the user is nervous, the emotion engine detects this and tells the interviewer that they are "nervous."

[0220] Through the above processing, the present invention eliminates gender bias and reflects the user's feelings, thereby achieving fair and objective interview evaluation.

[0221] The processing flow will be explained below.

[0222] Processing on the server

[0223] Step 1:

[0224] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[0225] Step 2:

[0226] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[0227] Step 3:

[0228] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[0229] Step 4:

[0230] The server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, analyzing voice characteristics such as tempo and pitch to identify the type of emotion.

[0231] Step 5:

[0232] The server sends the recognized emotion data to the interviewer's device, and transfers the emotion data in real time as additional information.

[0233] Processing on the device

[0234] Step 1:

[0235] The device receives a video stream from the user's camera, using OpenCV and other libraries to capture video frames from the camera.

[0236] Step 2:

[0237] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[0238] Step 3:

[0239] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[0240] Step 4:

[0241] The device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, for example, by analyzing facial movements and changes in facial features to identify emotions.

[0242] Step 5:

[0243] The device will then change the avatar's facial expression based on the recognized emotion, for example, making the avatar smile if the user is smiling.

[0244] Step 6:

[0245] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[0246] User processing

[0247] Step 1:

[0248] The user logs into the online interview system, accessing the system using a designated application or web portal.

[0249] Step 2:

[0250] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[0251] Step 3:

[0252] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[0253] Step 4:

[0254] The user's emotions are recognized in real time and communicated to the interviewer. For example, if the user is nervous, that information is displayed to the interviewer.

[0255] The above processing steps eliminate gender bias and allow for an online interview that reflects emotions, enabling a fair and objective evaluation.

[0256] Example 2

[0257] 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."

[0258] In conventional online interview systems, gender bias is often present in the user's voice and video, making it difficult to conduct fair evaluations. Furthermore, there is a lack of a way to recognize the user's emotions in real time and convey them to the interviewer, which means that interview evaluations tend to be subjective. There is a need to solve these issues and achieve fair and objective interview evaluations.

[0259] 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.

[0260] In this invention, the server includes means for receiving a user's voice stream and converting it to a gender-neutral voice, means for analyzing the voice stream and recognizing emotions, and means for outputting the recognized emotion data. This enables the elimination of gender bias and real-time recognition of user emotions. The server also includes means for acquiring a user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for analyzing the video stream and recognizing emotions from the user's facial expressions, means for changing the avatar's facial expressions based on the recognized emotions, and means for outputting the processed voice stream and video stream. This enables fair and objective interview evaluation.

[0261] An "audio stream" is a data flow that transmits a user's voice as digital data in real time.

[0262] A "gender-neutral voice" is a voice that is not biased towards a particular gender and makes it difficult for the listener to identify the gender.

[0263] A "means for transforming" is an algorithm or device that changes input data into a different format or characteristic.

[0264] A "video stream" is a data flow that transmits a user's video as digital data in real time.

[0265] "Sequential processing" is a method of processing continuously input data one by one.

[0266] A "means for detecting faces" is an algorithm or device for identifying and locating human faces within video data.

[0267] An "avatar" is a virtual character or image that is displayed in a digital space on behalf of a user.

[0268] A "superimposing means" is an algorithm or device for displaying separate images or data in an overlapping manner.

[0269] Analytical methods are methods for examining data in detail and extracting specific information or patterns from it.

[0270] "Means for recognizing emotions" refers to algorithms or devices for identifying a user's emotional state from audio or video data.

[0271] "Emotional data" is digital information about a user's perceived emotional state.

[0272] A "means for changing facial expressions" is an algorithm or device for changing the facial expression of an avatar in response to the emotional state of the user.

[0273] "Output means" refers to a method or device for providing processed data to an external entity.

[0274] This invention is an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[0275] Processing on the server

[0276] The server receives the user's voice stream and converts it into a gender-neutral voice. First, it receives the voice stream sent from the user's device. This voice stream is converted into a gender-neutral voice in real time using a dedicated voice conversion algorithm. For voice conversion, a Python (registered trademark)-based voice processing library (e.g., librosa) is used. The pitch and tone of the voice are adjusted to generate a voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's device.

[0277] The server then analyzes the audio stream and uses an emotion engine to recognize the user's emotions. The Microsoft® Azure® Emotion API is used for emotion recognition. The analyzed emotion data is sent to the interviewer's device in real time. For example, if the user speaks in an excited tone, the emotion engine recognizes this as "excitement" and displays the information on the interviewer's device.

[0278] Processing on the device

[0279] The device acquires the user's video stream. Each frame in the acquired video stream is processed sequentially in real time. The user's face is detected using a dedicated face detection algorithm. This process uses the OpenCV library. After face detection is complete, a neutral avatar image is superimposed on the detected face. Unity or Unreal Engine is used to generate and display the avatar.

[0280] The device also analyzes the video stream to recognize emotions from the user's facial expressions. Based on the recognized emotion, the avatar's facial expression changes. This emotion recognition again uses the Microsoft Azure Emotion API. For example, if the user is smiling, the emotion engine recognizes this as "joy," and the avatar is adjusted to smile as well.

[0281] User processing

[0282] The user logs into the online interview system and grants permission to use the camera and microphone. This permission allows the user's voice and video to be captured in real time. The user's voice is converted to a neutral voice, and a neutral avatar is displayed on the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the interviewer in real time.

[0283] As a concrete example, if a user enters the following prompt into an online interview system:

[0284] "Hello, I'm Taro Tanaka. I currently work as a data scientist, with particular expertise in machine learning and data analysis."

[0285] By entering this prompt, the system can accurately convey the user's gender-neutral voice and avatar to the interviewer.

[0286] From the above, it can be seen that this invention is a system that eliminates gender bias and reflects the user's emotions, thereby realizing fair and objective interview evaluation.

[0287] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0288] Step 1:

[0289] The server receives the audio stream sent from the user's device. Specifically, the audio data spoken by the user is sent to the server in digital form. This received data is input into the server's audio processing module. The output is audio data for conversion processing. For example, if the user says "hello," the audio data reaches the server.

[0290] Step 2:

[0291] The server converts the received audio stream into a gender-neutral voice. Here, it uses a Python-based audio processing library (e.g., librosa). It adjusts the pitch and tone of the input audio data to generate a gender-neutral voice that is not biased toward a specific gender. This converted audio data is then passed on to the next process. For example, a high-pitched male voice or a low-pitched female voice can be converted into a neutral voice.

[0292] Step 3:

[0293] The server sends the converted gender-neutral voice stream to the interviewer's device. The input is gender-neutral voice data, and the output is voice data sent to the interviewer's device. This allows the interviewer to receive voice without gender bias.

[0294] Step 4:

[0295] The server analyzes the audio stream and recognizes the user's emotion using an emotion engine (e.g., Microsoft Azure Emotion API). The input is the user's voice data, and the output is the recognized emotion data. For example, if the user speaks in an excited tone, the emotion engine will recognize this as "excited."

[0296] Step 5:

[0297] The server sends the recognized emotion data to the interviewer's device. The input is the recognized emotion data, and the output is the emotion information sent to the interviewer's device. This allows the interviewer to grasp the user's emotional state in real time.

[0298] Step 6:

[0299] The device acquires a video stream from the user's camera. The input is video data from the camera, and the output is video stream data. For example, when a user introduces themselves to the camera, the video is captured.

[0300] Step 7:

[0301] The device processes each frame in the captured video stream in real time and detects the user's face using a dedicated face detection algorithm (e.g., OpenCV). The input is the video stream data, and the output is the face coordinate data.

[0302] Step 8:

[0303] The device superimposes a neutral avatar image on the detected face. The input is face coordinate data and avatar image data, and the output is a video frame showing the avatar. This process is performed using Unity or Unreal Engine.

[0304] Step 9:

[0305] The device analyzes the video stream and recognizes emotions from the user's facial expressions. Emotion recognition uses the Microsoft Azure Emotion API. The input is the user's video stream data, and the output is the recognized emotion data. For example, if the user is smiling, the emotion engine recognizes this as "joy."

[0306] Step 10:

[0307] The device changes the facial expression of the avatar based on the recognized emotion. The input is the recognized emotion data, and the output is a video frame of the avatar with the changed facial expression. This allows the user's emotion to be visually conveyed to the interviewer.

[0308] Step 11:

[0309] A user logs in to the online interview system and grants permission to use the camera and microphone. The input is the user's login information, and the output is the access permission to the system. This allows the user's voice and video to be captured by the system.

[0310] Step 12:

[0311] The user's voice and video are captured in real time. The input is data from the camera and microphone, and the output is an audio stream and a video stream. For example, when a user introduces themselves, their voice and video are captured by the system.

[0312] Step 13:

[0313] The system converts the user's voice into a gender-neutral voice and displays a gender-neutral avatar on the screen. The input is the user's voice and video data, and the output is a gender-neutral voice and video data with an avatar.

[0314] Step 14:

[0315] The system recognizes the user's emotions in real time and conveys that information to the interviewer. The input is audio and video stream data, and the output is emotion data sent to the interviewer's device. For example, if the user is nervous, the emotion engine recognizes this as "nervous" and conveys that information to the interviewer.

[0316] (Application example 2)

[0317] 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."

[0318] Current online interview systems and other remote communication systems can be affected by gender bias. Furthermore, because users' emotions are not clearly conveyed, unfair evaluations may be made during interviews or customer service. It is necessary to solve these problems and achieve fair and objective communication.

[0319] 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 receiving a user's audio stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for recognizing emotions from the user's audio and video streams, means for changing the avatar's facial expression based on the recognized emotions, and means for outputting the processed audio and video streams. This eliminates gender bias and enables objective evaluation and fair communication that accurately reflects the user's emotions.

[0320] An "audio stream" is data that is continuously transmitted by converting the user's voice into a digital signal.

[0321] "Neutral voice" refers to a voice whose pitch and tone are adjusted to create a neutral voice that is not biased towards a specific gender.

[0322] A "video stream" is continuous video data captured from a user's camera.

[0323] "Sequential processing" refers to processing acquired data continuously in real time.

[0324] "Face detection" is the process of analyzing each frame in a video stream and identifying human faces.

[0325] An "avatar" is a virtual image of a person that is displayed on the screen in place of the user.

[0326] "Emotion recognition" refers to the process of analyzing and determining a user's emotional state from their audio and video streams.

[0327] "Changing facial expression" means dynamically changing the avatar's facial expression according to the recognized emotion.

[0328] "Output" refers to the action of displaying or transmitting processed data on another terminal or screen.

[0329] A "system" is a set of devices or programs in which multiple means operate in cooperation with each other.

[0330] The present invention provides a fair customer service experience in a virtual store based on an online interview system with gender-neutral voice conversion and emotion recognition functions. In this embodiment, a user, a terminal, and a server work together.

[0331] Processing on the server

[0332] The server receives the user's voice stream and converts it into a gender-neutral voice. This voice is generated by applying a voice conversion algorithm to ensure a gender-neutral voice. The server then uses an emotion engine to recognize emotions from the user's voice stream. This emotion data plays an important role in the customer service system of the virtual store.

[0333] Processing on the device

[0334] The device acquires a video stream from the user's camera and processes each frame in this video stream in real time. It uses a face detection algorithm such as OpenCV to detect the user's face in each frame. It then superimposes a neutral avatar image on the detected face. The device then analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions. It also has the ability to change the avatar's facial expression based on the recognized emotion.

[0335] User processing

[0336] The user logs into the virtual store customer service system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed in the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the store staff in real time. For example, if the user is excited about a product, the customer service system can use this information to make appropriate product introductions and suggestions, thereby increasing the user's desire to purchase.

[0337] Hardware and Software

[0338] Hardware: Camera, microphone, PC

[0339] Software: OpenCV, PyAudio, Emotion Recognizer (our own emotion recognition engine), Gender Neutral Voice (our own voice conversion module)

[0340] Specific examples

[0341] When a customer asks, "Tell me about this product," the system responds in a neutral voice, saying, "This product uses the latest technology and features X features." At the same time, if the customer is excited, the system notifies the store staff, who can then provide more detailed information or special offers.

[0342] Prompt Sentence Examples

[0343] "Give us an example of a process that detects faces in every video frame, overlays them with an androgynous avatar, and performs customer emotion recognition."

[0344] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0345] Step 1:

[0346] The user logs in to the virtual store customer service system and grants permission to use the camera and microphone.

[0347] Input: User's camera image and microphone audio

[0348] Action: Allow access to camera and microphone devices

[0349] Output: You can get the user's audio and video streams.

[0350] Step 2:

[0351] The device acquires the video stream from the camera and starts processing it sequentially in real time.

[0352] Input: Video stream from camera

[0353] Processing: Capture video stream frame by frame

[0354] Output: Individual video frames

[0355] Step 3:

[0356] The device uses OpenCV to detect the user's face in each video frame.

[0357] Input: Individual video frames

[0358] Processing: Apply OpenCV face detection algorithm to identify face area

[0359] Output: Video frames where faces are detected and face coordinate information

[0360] Step 4:

[0361] The device overlays a neutral avatar image onto the detected face.

[0362] Input: Coordinate information of face area

[0363] Processing: Overlay the avatar image onto the detected face area

[0364] Output: Video frame with avatar superimposed

[0365] Step 5:

[0366] The user's voice stream is also simultaneously captured by the terminal and sent to the server.

[0367] Input: Audio stream from microphone

[0368] Processing: Capture the audio stream and send it to the server

[0369] Output: The audio stream sent to the server

[0370] Step 6:

[0371] The server receives the audio stream and converts it to a neutral voice.

[0372] Input: The user's audio stream

[0373] Processing: Applying voice conversion algorithms to generate a neutral voice

[0374] Output: Neutral audio stream

[0375] Step 7:

[0376] The server utilizes an emotion engine to recognize the user's emotion from the audio stream.

[0377] Input: Neutral audio stream

[0378] Processing: Applying emotion engine to analyze emotional state from voice data

[0379] Output: Recognized emotion data

[0380] Step 8:

[0381] The recognized emotion data is transmitted to the terminal.

[0382] Input: Recognized emotion data

[0383] Processing: Emotion data is transferred to the device

[0384] Output: Emotion data sent to the device

[0385] Step 9:

[0386] The device analyzes the video stream and recognizes emotions from the user's facial expressions.

[0387] Input: Video frame

[0388] Processing: Apply video stream analysis and emotion engine to recognize emotions from facial expression data

[0389] Output: Emotion data from recognized facial expressions

[0390] Step 10:

[0391] The device changes the avatar's facial expression based on the recognized emotion.

[0392] Input: Emotion data recognized from voice and facial expressions

[0393] Processing: Dynamically change avatar facial expressions

[0394] Output: Video frames with an avatar superimposed with facial expressions corresponding to the emotion.

[0395] Step 11:

[0396] The processed audio and video streams are output in real time to shop staff in the virtual store.

[0397] Input: Neutral audio stream and video frames with avatars superimposed

[0398] Processing: Transfer of audio and video streams

[0399] Output: Audio and video streams displayed on shop staff's terminals

[0400] 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.

[0401] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0402] 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.

[0403] [Second embodiment]

[0404] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0405] 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.

[0406] 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).

[0407] 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.

[0408] 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.

[0409] 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).

[0410] 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.

[0411] 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.

[0412] 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.

[0413] 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.

[0414] 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.

[0415] 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."

[0416] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[0417] server

[0418] The server has the function of receiving the voice stream sent by the user. The received voice stream is converted into a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the input voice to generate a gender-neutral voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's terminal.

[0419] For example, if a user speaks in a high-pitched voice, the voice will be changed by the server to an appropriate pitch and transmitted to the interviewer in a neutral tone.

[0420] Terminal

[0421] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[0422] Specifically, a face detection algorithm identifies the position and size of the face within the frame, then resizes the avatar image to fit that size. The resized avatar is then overlaid on top of the original face image. This modified frame is displayed to both the user and the interviewer.

[0423] For example, if the user is smiling in the displayed frame, the avatar will be adjusted to appear smiling as well.

[0424] User

[0425] Users log in to the online interview system and grant permission to use the camera and microphone, which allows the system to capture their voice and video. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed on the video. This allows users to be interviewed without any bias regarding their gender.

[0426] For example, when a user introduces themselves in an online interview, their voice will be transmitted as a gender-neutral voice and a gender-neutral avatar will be displayed on the video screen. The interviewer will see and hear this gender-neutral voice and avatar, allowing the interview to proceed without being influenced by gender bias.

[0427] This invention allows interviewers to evaluate candidates without gender bias, thereby realizing fair and objective interviews.

[0428] The processing flow will be explained below.

[0429] Processing on the server

[0430] Step 1:

[0431] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[0432] Step 2:

[0433] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[0434] Step 3:

[0435] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[0436] Processing on the device

[0437] Step 1:

[0438] The device receives a video stream from the user's camera and uses OpenCV and other libraries to capture video frames from the camera.

[0439] Step 2:

[0440] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[0441] Step 3:

[0442] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[0443] Step 4:

[0444] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[0445] User processing

[0446] Step 1:

[0447] The user logs into the online interview system, accessing the system using a designated application or web portal.

[0448] Step 2:

[0449] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[0450] Step 3:

[0451] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[0452] The above processing steps enable online interviews that are free of gender bias.

[0453] Example 1

[0454] 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."

[0455] In conventional online interview systems, gender bias can affect evaluations. A particular problem is that interviewers may unconsciously have preconceived notions based on the candidate's gender. This makes it difficult to conduct fair and objective evaluations, and it is difficult to accurately determine the candidate's true abilities and aptitude.

[0456] 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.

[0457] In this invention, the server includes means for receiving the user's voice data and converting it into a gender-neutral voice, means for acquiring the user's video data and sequentially processing it to detect the user's face, and means for superimposing a gender-neutral virtual character on the detected face, thereby eliminating gender bias and enabling fair and objective interviews.

[0458] A "user" is a person who uses the online interview system, accesses the system, and provides their own audio and video.

[0459] "Audio Data" refers to audio information collected through the user's microphone, which is sent to the server for processing.

[0460] "Gender-neutral voice" refers to voices with pitch and tone adjusted so that they are not associated with a specific gender.

[0461] "Video data" refers to video information collected through a user's camera, material that is processed frame by frame.

[0462] "Sequential processing" refers to a method of processing input data sequentially in real time.

[0463] "Face detection" refers to the process of identifying the location and size of a user's face from frames within video data.

[0464] "Virtual characters" are neutral images or animations that are superimposed on the user's face and are designed to avoid any gender-related associations.

[0465] "Superimposition" is the process of superimposing one image or object on top of another.

[0466] "Gender bias" is a term that refers to preconceived notions or prejudices against a particular gender, and is an unfair factor that influences evaluations and judgments.

[0467] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[0468] server

[0469] The server has hardware and software to receive the voice data sent by the user. After receiving the voice data, the server uses the Vocoder algorithm to adjust the pitch and tone of the voice and convert it into a neutral voice. This converted voice data is then sent to the interviewer's device in real time.

[0470] For example, if a user provides a normal speaking voice, the voice is analyzed by the server and normalized to a standard pitch and tone.

[0471] Terminal

[0472] The device is equipped with hardware and software to capture video data from the user's camera. The captured video data is then processed sequentially using a face detection algorithm, such as OpenCV or Dlib, to identify the location and size of the user's face and overlay a gender-neutral virtual character on it.

[0473] This superimposed video data is displayed in real time to both the user and the interviewer. For example, when the user speaks with a smile, the virtual character is adjusted so that it also appears smiling based on that facial expression.

[0474] User

[0475] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured and processed on the server and device. The user's voice is converted to a gender-neutral voice, and a gender-neutral virtual character is superimposed on the video. This system allows the user to be interviewed without any gender bias.

[0476] For example, when a user introduces themselves in an online interview, their voice is converted to a gender-neutral voice on the server side, and a gender-neutral virtual character is displayed on the video screen. Through this gender-neutral voice and virtual character, the interviewer can conduct the interview without being influenced by gender bias.

[0477] Examples of prompt statements

[0478] Explain an online interview system to eliminate gender bias. Describe in detail the specific operation and examples of each component (server, terminal, user).

[0479] This invention is expected to enable fair and objective interviews and improve the accuracy of evaluations.

[0480] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0481] Step 1:

[0482] The user logs in to the online interview system. They enter their user ID and password on the login screen to access the system. The system then authenticates the user and sets up the system to start the interview session. The input is the user's authentication information, and the output is the system access rights after successful login.

[0483] Step 2:

[0484] The user grants permission to access the camera and microphone. The system displays a pop-up message, and the user must click "Allow" to access the device. The input is the user's permission, and the output is the connection status of the camera and microphone.

[0485] Step 3:

[0486] The user's device acquires video data from the camera and audio data from the microphone. These data are sent to the server as input data. For example, the camera captures video at 30 frames per second, and the microphone captures audio in real time. The input is the video and audio data, and the output is the data sent to the server.

[0487] Step 4:

[0488] The server receives the transmitted voice data and converts it into a gender-neutral voice using the Vocoder algorithm, adjusting the pitch and tone of the voice in real time to eliminate gender cues. The input is the original voice data, and the output is the converted gender-neutral voice data.

[0489] Step 5:

[0490] The converted neutral voice data is sent from the server to the interviewer's device in real time. The input is neutral voice data, and the output is the data sent to the interviewer's device. For example, if the user speaks in a high-pitched voice, the voice will be converted to a flat, neutral voice and delivered to the interviewer.

[0491] Step 6:

[0492] The device processes the video data acquired from the camera in real time and applies a face detection algorithm (e.g., OpenCV or Dlib) to identify the position and size of the user's face. The input is the original video data, and the output is the position and size information of the face.

[0493] Step 7:

[0494] Based on the detected face position and size, an image processing algorithm generates a gender-neutral virtual character and superimposes it on the original face image. The input is face position and size information, and the output is video data with the virtual character superimposed. For example, if a user is smiling while talking, the gender-neutral virtual character will also be displayed smiling.

[0495] Step 8:

[0496] The processed video data is displayed in real time to both the user and the interviewer. The input is video data with a virtual character superimposed, and the output is a video feed displayed on the user's and interviewer's devices.

[0497] Step 9:

[0498] During the interview, the interviewer asks the user questions through a neutral voice and virtual character, and the user answers. This repeated process is carried out in real time on the server and the terminal. The input is the interviewer's questions and the user's answers, and the output is real-time alternating communication between them.

[0499] Step 10:

[0500] After the interview is over, the user presses the "End" button to end the interview. This causes the server to stop processing the audio and video streams and end the session. The input is the end command, and the output is the stopped state of the audio and video streams.

[0501] (Application example 1)

[0502] 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."

[0503] Audio and video neutralization technology was invented to eliminate gender bias in online interview systems. Similarly, various voice operations and interactions are increasingly required in autonomous vehicles, where gender bias also presents a problem. Conventional in-vehicle interactive systems can contain unconscious gender bias in voice assistants and display displays, resulting in a lack of fairness for passengers and drivers.

[0504] 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.

[0505] In this invention, the server includes means for receiving a user's voice stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for outputting the processed voice stream and video stream to an interactive system of the autonomous vehicle, means for outputting the user's gender-neutral voice to an in-vehicle speaker, and means for displaying the gender-neutral avatar on an in-vehicle display using face detection, thereby eliminating gender bias in voice operations and display views in the autonomous vehicle and enabling fair and neutral interactions.

[0506] "User" refers to a person who uses a system or application.

[0507] An "audio stream" refers to a continuously transmitted stream of audio data.

[0508] "Gender-neutral voice" refers to a voice that has vocal characteristics that are not biased towards a specific gender.

[0509] A "video stream" refers to a continuously transmitted stream of video data.

[0510] "Face detection" refers to the process of recognizing the location and size of a face in a particular frame within a video stream.

[0511] A "gender-neutral avatar" refers to a virtual character whose visual elements are not gender-biased.

[0512] An "autonomous vehicle" refers to a vehicle that drives autonomously without the need for driver intervention.

[0513] An "interactive system" refers to a system that allows two-way communication with the user.

[0514] "In-vehicle speaker" refers to an audio output device installed inside a vehicle.

[0515] "Server" refers to a computer system that provides resources and processes data over a network.

[0516] "Sequential processing" refers to a method of processing data one by one in succession.

[0517] An "algorithm" refers to a method of solving a problem using specific computational procedures.

[0518] "Display" refers to a device or screen for displaying images or videos.

[0519] The present invention provides a technology for eliminating gender bias in interactive systems within autonomous vehicles. Specific embodiments of this system are described below.

[0520] Hardware and software used

[0521] 1. Hardware

[0522] Server: Located remotely and responsible for processing the audio and video streams.

[0523] Vehicle microphone: A device for capturing the user's voice stream.

[0524] Dash camera: A device for capturing a user's video stream.

[0525] Car speaker: A device for outputting converted neutral audio.

[0526] In-car display: A screen for displaying a gender-neutral avatar in the car.

[0527] 2. Software

[0528] Speech recognition engine (e.g., Google Speech-to-Text API, Amazon Transcribe): Recognizes the user's audio stream and converts it into text.

[0529] Voice conversion algorithms (e.g., Voice Conversion NN(VCNN)): convert a voice stream into a neutral voice.

[0530] Face detection libraries (e.g. OpenCV): Detect faces in the captured video stream.

[0531] Avatar generation engine (e.g. Unity): Generates gender-neutral avatars based on face detection results.

[0532] User interface frameworks (e.g., React Native, Flutter): for building the front-end of in-car interactive systems.

[0533] Server Procedure

[0534] 1. Audio capture and conversion

[0535] The server receives the audio stream transmitted from the vehicle microphone.

[0536] The received audio stream is converted to a gender-neutral voice using a voice conversion algorithm, which adjusts pitch and tone to eliminate gender bias.

[0537] The converted neutral voice is output to the car's speakers.

[0538] 2. Video Stream Processing

[0539] The server receives the video stream transmitted from the in-car camera.

[0540] Each frame in the video stream is processed in real time using a face detection library.

[0541] A gender-neutral avatar generated by an avatar generation engine is superimposed on the detected face.

[0542] The generated avatar is displayed on the in-car display.

[0543] Specific examples

[0544] Audio capture and conversion examples

[0545] When a user says "Start navigation" to the in-car assistant, the server converts this speech into a gender-neutral voice in real time. The in-car speaker plays "Start navigation" in a consistent, gender-neutral tone. This real-time conversion is achieved using a voice conversion algorithm.

[0546] Video stream processing example

[0547] When a user enters a vehicle, the dashcam starts capturing the video stream. The server detects faces in each frame and displays a neutral avatar. For example, if the user is smiling, the avatar will adjust to smile as well.

[0548] Example prompts

[0549] "Convert to gender-neutral voice. This voice is a command for your in-car assistant."

[0550] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0551] Step 1: Capture the audio stream

[0552] The user speaks a voice command to the in-car assistant.

[0553] Input: User's voice captured by the vehicle's microphone.

[0554] Processing: The vehicle microphone sends the audio stream to the server.

[0555] Output: Real-time audio stream sent to the server.

[0556] Step 2: Neutralizing the Audio Stream

[0557] The server inputs the received audio stream into a voice conversion algorithm.

[0558] Input: The audio stream sent to the server.

[0559] Processing: A voice conversion algorithm (e.g., Voice Conversion NN (VCNN)) adjusts the pitch and tone of the voice to convert it into a neutral voice.

[0560] Output: Neutral converted audio stream.

[0561] Step 3: Output audio

[0562] The server sends a neutral audio stream to the car's speakers.

[0563] Input: Neutral converted audio stream.

[0564] Processing: A neutral audio stream is sent to the car speakers.

[0565] Output: Neutral voice played through the car speakers.

[0566] Step 4: Capturing the video stream

[0567] While the user is inside the car, the on-board camera captures a video stream.

[0568] Input: Video stream captured by the dashcam.

[0569] Processing: The video stream is sent to the server.

[0570] Output: Real-time video stream sent to the server.

[0571] Step 5: Face detection

[0572] The server inputs the received video stream into a face detection algorithm.

[0573] Input: The video stream sent to the server.

[0574] Processing: A face detection algorithm (e.g. OpenCV) detects the position and size of the user's face in each frame.

[0575] Output: Face position and size information for each frame.

[0576] Step 6: Avatar generation

[0577] The server inputs the face detection information into an avatar generation engine to generate a gender-neutral avatar.

[0578] Input: Face position and size information.

[0579] Processing: An avatar generation engine (e.g., Unity) generates a gender-neutral avatar based on the detected facial features and overlays it on each frame.

[0580] Output: A frame with an androgynous avatar superimposed.

[0581] Step 7: Output the video stream

[0582] The server transmits the video stream with the avatar superimposed on it to an in-car display.

[0583] Input: A video frame with an androgynous avatar superimposed on it.

[0584] Processing: The video stream is sent to the in-car display.

[0585] Output: A video stream with a gender-neutral avatar displayed on the in-car display.

[0586] 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.

[0587] The present invention relates to an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[0588] Processing on the server

[0589] The server receives the voice stream sent from the user's device. The received voice stream is converted to a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the voice to create a gender-neutral voice. The converted voice stream is then sent to the interviewer's device.

[0590] Furthermore, the server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, and this recognized emotion data is also sent to the interviewer's device.

[0591] For example, if a user speaks in an excited tone, the emotion engine will recognize this as "excited" and convey that information to the interviewer.

[0592] Processing on the device

[0593] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[0594] Furthermore, the device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, and changes the avatar's facial expression based on the recognized emotions.

[0595] For example, if the user is smiling, the emotion engine will recognize this as "happiness" and adjust the avatar to smile as well.

[0596] User processing

[0597] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the system to capture the user's voice and video. The user's voice is converted to a neutral voice and a neutral avatar is displayed on the video. Furthermore, the system recognizes the user's emotions and relays this information to the interviewer in real time.

[0598] For example, when a user introduces themselves in an online interview, their voice is conveyed as a neutral voice and a neutral avatar is displayed on the video screen. If the user is nervous, the emotion engine detects this and tells the interviewer that they are "nervous."

[0599] Through the above processing, the present invention eliminates gender bias and reflects the user's feelings, thereby achieving fair and objective interview evaluation.

[0600] The processing flow will be explained below.

[0601] Processing on the server

[0602] Step 1:

[0603] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[0604] Step 2:

[0605] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[0606] Step 3:

[0607] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[0608] Step 4:

[0609] The server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, analyzing voice characteristics such as tempo and pitch to identify the type of emotion.

[0610] Step 5:

[0611] The server sends the recognized emotion data to the interviewer's device, and transfers the emotion data in real time as additional information.

[0612] Processing on the device

[0613] Step 1:

[0614] The device receives a video stream from the user's camera, using OpenCV and other libraries to capture video frames from the camera.

[0615] Step 2:

[0616] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[0617] Step 3:

[0618] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[0619] Step 4:

[0620] The device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, for example, by analyzing facial movements and changes in facial features to identify emotions.

[0621] Step 5:

[0622] The device will then change the avatar's facial expression based on the recognized emotion, for example, making the avatar smile if the user is smiling.

[0623] Step 6:

[0624] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[0625] User processing

[0626] Step 1:

[0627] The user logs into the online interview system, accessing the system using a designated application or web portal.

[0628] Step 2:

[0629] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[0630] Step 3:

[0631] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[0632] Step 4:

[0633] The user's emotions are recognized in real time and communicated to the interviewer. For example, if the user is nervous, that information is displayed to the interviewer.

[0634] The above processing steps eliminate gender bias and allow for an online interview that reflects emotions, enabling a fair and objective evaluation.

[0635] Example 2

[0636] 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."

[0637] In conventional online interview systems, gender bias is often present in the user's voice and video, making it difficult to conduct fair evaluations. Furthermore, there is a lack of a way to recognize the user's emotions in real time and convey them to the interviewer, which means that interview evaluations tend to be subjective. There is a need to solve these issues and achieve fair and objective interview evaluations.

[0638] 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.

[0639] In this invention, the server includes means for receiving a user's voice stream and converting it to a gender-neutral voice, means for analyzing the voice stream and recognizing emotions, and means for outputting the recognized emotion data. This enables the elimination of gender bias and real-time recognition of user emotions. The server also includes means for acquiring a user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for analyzing the video stream and recognizing emotions from the user's facial expressions, means for changing the avatar's facial expressions based on the recognized emotions, and means for outputting the processed voice stream and video stream. This enables fair and objective interview evaluation.

[0640] An "audio stream" is a data flow that transmits a user's voice as digital data in real time.

[0641] A "gender-neutral voice" is a voice that is not biased towards a particular gender and makes it difficult for the listener to identify the gender.

[0642] A "means for transforming" is an algorithm or device that changes input data into a different format or characteristic.

[0643] A "video stream" is a data flow that transmits a user's video as digital data in real time.

[0644] "Sequential processing" is a method of processing continuously input data one by one.

[0645] A "means for detecting faces" is an algorithm or device for identifying and locating human faces within video data.

[0646] An "avatar" is a virtual character or image that is displayed in a digital space on behalf of a user.

[0647] A "superimposing means" is an algorithm or device for displaying separate images or data in an overlapping manner.

[0648] Analytical methods are methods for examining data in detail and extracting specific information or patterns from it.

[0649] "Means for recognizing emotions" refers to algorithms or devices for identifying a user's emotional state from audio or video data.

[0650] "Emotional data" is digital information about a user's perceived emotional state.

[0651] A "means for changing facial expressions" is an algorithm or device for changing the facial expression of an avatar in response to the emotional state of the user.

[0652] "Output means" refers to a method or device for providing processed data to an external entity.

[0653] This invention is an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[0654] Processing on the server

[0655] The server receives the user's voice stream and converts it into a gender-neutral voice. First, it receives the voice stream sent from the user's device. This voice stream is converted into a gender-neutral voice in real time using a dedicated voice conversion algorithm. For voice conversion, a Python-based voice processing library (e.g., librosa) is used. The pitch and tone of the voice are adjusted to produce a voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's device.

[0656] The server then analyzes the audio stream and uses an emotion engine to recognize the user's emotions. The emotion recognition uses the Microsoft Azure Emotion API. The analyzed emotion data is sent to the interviewer's device in real time. For example, if the user speaks in an excited tone, the emotion engine recognizes this as "excitement," and the information is displayed on the interviewer's device.

[0657] Processing on the device

[0658] The device acquires the user's video stream. Each frame in the acquired video stream is processed sequentially in real time. The user's face is detected using a dedicated face detection algorithm. This process uses the OpenCV library. After face detection is complete, a neutral avatar image is superimposed on the detected face. Unity or Unreal Engine is used to generate and display the avatar.

[0659] The device also analyzes the video stream to recognize emotions from the user's facial expressions. Based on the recognized emotion, the avatar's facial expression changes. This emotion recognition again uses the Microsoft Azure Emotion API. For example, if the user is smiling, the emotion engine recognizes this as "joy," and the avatar is adjusted to smile as well.

[0660] User processing

[0661] The user logs into the online interview system and grants permission to use the camera and microphone. This permission allows the user's voice and video to be captured in real time. The user's voice is converted to a neutral voice, and a neutral avatar is displayed on the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the interviewer in real time.

[0662] As a concrete example, if a user enters the following prompt into an online interview system:

[0663] "Hello, I'm Taro Tanaka. I currently work as a data scientist, with particular expertise in machine learning and data analysis."

[0664] By entering this prompt, the system can accurately convey the user's gender-neutral voice and avatar to the interviewer.

[0665] From the above, it can be seen that this invention is a system that eliminates gender bias and reflects the user's emotions, thereby realizing fair and objective interview evaluation.

[0666] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0667] Step 1:

[0668] The server receives the audio stream sent from the user's device. Specifically, the audio data spoken by the user is sent to the server in digital form. This received data is input into the server's audio processing module. The output is audio data for conversion processing. For example, if the user says "hello," the audio data reaches the server.

[0669] Step 2:

[0670] The server converts the received audio stream into a gender-neutral voice. Here, it uses a Python-based audio processing library (e.g., librosa). It adjusts the pitch and tone of the input audio data to generate a gender-neutral voice that is not biased toward a specific gender. This converted audio data is then passed on to the next process. For example, a high-pitched male voice or a low-pitched female voice can be converted into a neutral voice.

[0671] Step 3:

[0672] The server sends the converted gender-neutral voice stream to the interviewer's device. The input is gender-neutral voice data, and the output is voice data sent to the interviewer's device. This allows the interviewer to receive voice without gender bias.

[0673] Step 4:

[0674] The server analyzes the audio stream and recognizes the user's emotion using an emotion engine (e.g., Microsoft Azure Emotion API). The input is the user's voice data, and the output is the recognized emotion data. For example, if the user speaks in an excited tone, the emotion engine will recognize this as "excited."

[0675] Step 5:

[0676] The server sends the recognized emotion data to the interviewer's device. The input is the recognized emotion data, and the output is the emotion information sent to the interviewer's device. This allows the interviewer to grasp the user's emotional state in real time.

[0677] Step 6:

[0678] The device acquires a video stream from the user's camera. The input is video data from the camera, and the output is video stream data. For example, when a user introduces themselves to the camera, the video is captured.

[0679] Step 7:

[0680] The device processes each frame in the captured video stream in real time and detects the user's face using a dedicated face detection algorithm (e.g., OpenCV). The input is the video stream data, and the output is the face coordinate data.

[0681] Step 8:

[0682] The device superimposes a neutral avatar image on the detected face. The input is face coordinate data and avatar image data, and the output is a video frame showing the avatar. This process is performed using Unity or Unreal Engine.

[0683] Step 9:

[0684] The device analyzes the video stream and recognizes emotions from the user's facial expressions. Emotion recognition uses the Microsoft Azure Emotion API. The input is the user's video stream data, and the output is the recognized emotion data. For example, if the user is smiling, the emotion engine recognizes this as "joy."

[0685] Step 10:

[0686] The device changes the facial expression of the avatar based on the recognized emotion. The input is the recognized emotion data, and the output is a video frame of the avatar with the changed facial expression. This allows the user's emotion to be visually conveyed to the interviewer.

[0687] Step 11:

[0688] A user logs in to the online interview system and grants permission to use the camera and microphone. The input is the user's login information, and the output is the access permission to the system. This allows the user's voice and video to be captured by the system.

[0689] Step 12:

[0690] The user's voice and video are captured in real time. The input is data from the camera and microphone, and the output is an audio stream and a video stream. For example, when a user introduces themselves, their voice and video are captured by the system.

[0691] Step 13:

[0692] The system converts the user's voice into a gender-neutral voice and displays a gender-neutral avatar on the screen. The input is the user's voice and video data, and the output is a gender-neutral voice and video data with an avatar.

[0693] Step 14:

[0694] The system recognizes the user's emotions in real time and conveys that information to the interviewer. The input is audio and video stream data, and the output is emotion data sent to the interviewer's device. For example, if the user is nervous, the emotion engine recognizes this as "nervous" and conveys that information to the interviewer.

[0695] (Application example 2)

[0696] 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."

[0697] Current online interview systems and other remote communication systems can be affected by gender bias. Furthermore, because users' emotions are not clearly conveyed, unfair evaluations may be made during interviews or customer service. It is necessary to solve these problems and achieve fair and objective communication.

[0698] 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 receiving a user's audio stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for recognizing emotions from the user's audio and video streams, means for changing the avatar's facial expression based on the recognized emotions, and means for outputting the processed audio and video streams. This eliminates gender bias and enables objective evaluation and fair communication that accurately reflects the user's emotions.

[0699] An "audio stream" is data that is continuously transmitted by converting the user's voice into a digital signal.

[0700] "Neutral voice" refers to a voice whose pitch and tone are adjusted to create a neutral voice that is not biased towards a specific gender.

[0701] A "video stream" is continuous video data captured from a user's camera.

[0702] "Sequential processing" refers to processing acquired data continuously in real time.

[0703] "Face detection" is the process of analyzing each frame in a video stream and identifying human faces.

[0704] An "avatar" is a virtual image of a person that is displayed on the screen in place of the user.

[0705] "Emotion recognition" refers to the process of analyzing and determining a user's emotional state from their audio and video streams.

[0706] "Changing facial expression" means dynamically changing the avatar's facial expression according to the recognized emotion.

[0707] "Output" refers to the action of displaying or transmitting processed data on another terminal or screen.

[0708] A "system" is a set of devices or programs in which multiple means operate in cooperation with each other.

[0709] The present invention provides a fair customer service experience in a virtual store based on an online interview system with gender-neutral voice conversion and emotion recognition functions. In this embodiment, a user, a terminal, and a server work together.

[0710] Processing on the server

[0711] The server receives the user's voice stream and converts it into a gender-neutral voice. This voice is generated by applying a voice conversion algorithm to ensure a gender-neutral voice. The server then uses an emotion engine to recognize emotions from the user's voice stream. This emotion data plays an important role in the customer service system of the virtual store.

[0712] Processing on the device

[0713] The device acquires a video stream from the user's camera and processes each frame in this video stream in real time. It uses a face detection algorithm such as OpenCV to detect the user's face in each frame. It then superimposes a neutral avatar image on the detected face. The device then analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions. It also has the ability to change the avatar's facial expression based on the recognized emotion.

[0714] User processing

[0715] The user logs into the virtual store customer service system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed in the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the store staff in real time. For example, if the user is excited about a product, the customer service system can use this information to make appropriate product introductions and suggestions, thereby increasing the user's desire to purchase.

[0716] Hardware and Software

[0717] Hardware: Camera, microphone, PC

[0718] Software: OpenCV, PyAudio, Emotion Recognizer (our own emotion recognition engine), Gender Neutral Voice (our own voice conversion module)

[0719] Specific examples

[0720] When a customer asks, "Tell me about this product," the system responds in a neutral voice, saying, "This product uses the latest technology and features X features." At the same time, if the customer is excited, the system notifies the store staff, who can then provide more detailed information or special offers.

[0721] Prompt Sentence Examples

[0722] "Give us an example of a process that detects faces in every video frame, overlays them with an androgynous avatar, and performs customer emotion recognition."

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

[0724] Step 1:

[0725] The user logs in to the virtual store customer service system and grants permission to use the camera and microphone.

[0726] Input: User's camera image and microphone audio

[0727] Action: Allow access to camera and microphone devices

[0728] Output: You can get the user's audio and video streams.

[0729] Step 2:

[0730] The device acquires the video stream from the camera and starts processing it sequentially in real time.

[0731] Input: Video stream from camera

[0732] Processing: Capture video stream frame by frame

[0733] Output: Individual video frames

[0734] Step 3:

[0735] The device uses OpenCV to detect the user's face in each video frame.

[0736] Input: Individual video frames

[0737] Processing: Apply OpenCV face detection algorithm to identify face area

[0738] Output: Video frames where faces are detected and face coordinate information

[0739] Step 4:

[0740] The device overlays a neutral avatar image onto the detected face.

[0741] Input: Coordinate information of face area

[0742] Processing: Overlay the avatar image onto the detected face area

[0743] Output: Video frame with avatar superimposed

[0744] Step 5:

[0745] The user's voice stream is also simultaneously captured by the terminal and sent to the server.

[0746] Input: Audio stream from microphone

[0747] Processing: Capture the audio stream and send it to the server

[0748] Output: The audio stream sent to the server

[0749] Step 6:

[0750] The server receives the audio stream and converts it to a neutral voice.

[0751] Input: The user's audio stream

[0752] Processing: Applying voice conversion algorithms to generate a neutral voice

[0753] Output: Neutral audio stream

[0754] Step 7:

[0755] The server utilizes an emotion engine to recognize the user's emotion from the audio stream.

[0756] Input: Neutral audio stream

[0757] Processing: Applying emotion engine to analyze emotional state from voice data

[0758] Output: Recognized emotion data

[0759] Step 8:

[0760] The recognized emotion data is transmitted to the terminal.

[0761] Input: Recognized emotion data

[0762] Processing: Emotion data is transferred to the device

[0763] Output: Emotion data sent to the device

[0764] Step 9:

[0765] The device analyzes the video stream and recognizes emotions from the user's facial expressions.

[0766] Input: Video frame

[0767] Processing: Apply video stream analysis and emotion engine to recognize emotions from facial expression data

[0768] Output: Emotion data from recognized facial expressions

[0769] Step 10:

[0770] The device changes the avatar's facial expression based on the recognized emotion.

[0771] Input: Emotion data recognized from voice and facial expressions

[0772] Processing: Dynamically change avatar facial expressions

[0773] Output: Video frames with an avatar superimposed with facial expressions corresponding to the emotion.

[0774] Step 11:

[0775] The processed audio and video streams are output in real time to shop staff in the virtual store.

[0776] Input: Neutral audio stream and video frames with avatars superimposed

[0777] Processing: Transfer of audio and video streams

[0778] Output: Audio and video streams displayed on shop staff's terminals

[0779] 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.

[0780] 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.

[0781] 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.

[0782] [Third embodiment]

[0783] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0784] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0785] 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).

[0786] 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.

[0787] 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.

[0788] 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).

[0789] 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.

[0790] 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.

[0791] 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.

[0792] 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.

[0793] 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.

[0794] 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."

[0795] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[0796] server

[0797] The server has the function of receiving the voice stream sent by the user. The received voice stream is converted into a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the input voice to generate a gender-neutral voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's terminal.

[0798] For example, if a user speaks in a high-pitched voice, the voice will be changed by the server to an appropriate pitch and transmitted to the interviewer in a neutral tone.

[0799] Terminal

[0800] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[0801] Specifically, a face detection algorithm identifies the position and size of the face within the frame, then resizes the avatar image to fit that size. The resized avatar is then overlaid on top of the original face image. This modified frame is displayed to both the user and the interviewer.

[0802] For example, if the user is smiling in the displayed frame, the avatar will be adjusted to appear smiling as well.

[0803] User

[0804] Users log in to the online interview system and grant permission to use the camera and microphone, which allows the system to capture their voice and video. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed on the video. This allows users to be interviewed without any bias regarding their gender.

[0805] For example, when a user introduces themselves in an online interview, their voice will be transmitted as a gender-neutral voice and a gender-neutral avatar will be displayed on the video screen. The interviewer will see and hear this gender-neutral voice and avatar, allowing the interview to proceed without being influenced by gender bias.

[0806] This invention allows interviewers to evaluate candidates without gender bias, thereby realizing fair and objective interviews.

[0807] The processing flow will be explained below.

[0808] Processing on the server

[0809] Step 1:

[0810] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[0811] Step 2:

[0812] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[0813] Step 3:

[0814] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[0815] Processing on the device

[0816] Step 1:

[0817] The device receives a video stream from the user's camera and uses OpenCV and other libraries to capture video frames from the camera.

[0818] Step 2:

[0819] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[0820] Step 3:

[0821] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[0822] Step 4:

[0823] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[0824] User processing

[0825] Step 1:

[0826] The user logs into the online interview system, accessing the system using a designated application or web portal.

[0827] Step 2:

[0828] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[0829] Step 3:

[0830] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[0831] The above processing steps enable online interviews that are free of gender bias.

[0832] Example 1

[0833] 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."

[0834] In conventional online interview systems, gender bias can affect evaluations. A particular problem is that interviewers may unconsciously have preconceived notions based on the candidate's gender. This makes it difficult to conduct fair and objective evaluations, and it is difficult to accurately determine the candidate's true abilities and aptitude.

[0835] 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.

[0836] In this invention, the server includes means for receiving the user's voice data and converting it into a gender-neutral voice, means for acquiring the user's video data and sequentially processing it to detect the user's face, and means for superimposing a gender-neutral virtual character on the detected face, thereby eliminating gender bias and enabling fair and objective interviews.

[0837] A "user" is a person who uses the online interview system, accesses the system, and provides their own audio and video.

[0838] "Audio Data" refers to audio information collected through the user's microphone, which is sent to the server for processing.

[0839] "Gender-neutral voice" refers to voices with pitch and tone adjusted so that they are not associated with a specific gender.

[0840] "Video data" refers to video information collected through a user's camera, material that is processed frame by frame.

[0841] "Sequential processing" refers to a method of processing input data sequentially in real time.

[0842] "Face detection" refers to the process of identifying the location and size of a user's face from frames within video data.

[0843] "Virtual characters" are neutral images or animations that are superimposed on the user's face and are designed to avoid any gender-related associations.

[0844] "Superimposition" is the process of superimposing one image or object on top of another.

[0845] "Gender bias" is a term that refers to preconceived notions or prejudices against a particular gender, and is an unfair factor that influences evaluations and judgments.

[0846] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[0847] server

[0848] The server has hardware and software to receive the voice data sent by the user. After receiving the voice data, the server uses the Vocoder algorithm to adjust the pitch and tone of the voice and convert it into a neutral voice. This converted voice data is then sent to the interviewer's device in real time.

[0849] For example, if a user provides a normal speaking voice, the voice is analyzed by the server and normalized to a standard pitch and tone.

[0850] Terminal

[0851] The device is equipped with hardware and software to capture video data from the user's camera. The captured video data is then processed sequentially using a face detection algorithm, such as OpenCV or Dlib, to identify the location and size of the user's face and overlay a gender-neutral virtual character on it.

[0852] This superimposed video data is displayed in real time to both the user and the interviewer. For example, when the user speaks with a smile, the virtual character is adjusted so that it also appears smiling based on that facial expression.

[0853] User

[0854] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured and processed on the server and device. The user's voice is converted to a gender-neutral voice, and a gender-neutral virtual character is superimposed on the video. This system allows the user to be interviewed without any gender bias.

[0855] For example, when a user introduces themselves in an online interview, their voice is converted to a gender-neutral voice on the server side, and a gender-neutral virtual character is displayed on the video screen. Through this gender-neutral voice and virtual character, the interviewer can conduct the interview without being influenced by gender bias.

[0856] Examples of prompt statements

[0857] Explain an online interview system to eliminate gender bias. Describe in detail the specific operation and examples of each component (server, terminal, user).

[0858] This invention is expected to enable fair and objective interviews and improve the accuracy of evaluations.

[0859] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0860] Step 1:

[0861] The user logs in to the online interview system. They enter their user ID and password on the login screen to access the system. The system then authenticates the user and sets up the system to start the interview session. The input is the user's authentication information, and the output is the system access rights after successful login.

[0862] Step 2:

[0863] The user grants permission to access the camera and microphone. The system displays a pop-up message, and the user must click "Allow" to access the device. The input is the user's permission, and the output is the connection status of the camera and microphone.

[0864] Step 3:

[0865] The user's device acquires video data from the camera and audio data from the microphone. These data are sent to the server as input data. For example, the camera captures video at 30 frames per second, and the microphone captures audio in real time. The input is the video and audio data, and the output is the data sent to the server.

[0866] Step 4:

[0867] The server receives the transmitted voice data and converts it into a gender-neutral voice using the Vocoder algorithm, adjusting the pitch and tone of the voice in real time to eliminate gender cues. The input is the original voice data, and the output is the converted gender-neutral voice data.

[0868] Step 5:

[0869] The converted neutral voice data is sent from the server to the interviewer's device in real time. The input is neutral voice data, and the output is the data sent to the interviewer's device. For example, if the user speaks in a high-pitched voice, the voice will be converted to a flat, neutral voice and delivered to the interviewer.

[0870] Step 6:

[0871] The device processes the video data acquired from the camera in real time and applies a face detection algorithm (e.g., OpenCV or Dlib) to identify the position and size of the user's face. The input is the original video data, and the output is the position and size information of the face.

[0872] Step 7:

[0873] Based on the detected face position and size, an image processing algorithm generates a gender-neutral virtual character and superimposes it on the original face image. The input is face position and size information, and the output is video data with the virtual character superimposed. For example, if a user is smiling while talking, the gender-neutral virtual character will also be displayed smiling.

[0874] Step 8:

[0875] The processed video data is displayed in real time to both the user and the interviewer. The input is video data with a virtual character superimposed, and the output is a video feed displayed on the user's and interviewer's devices.

[0876] Step 9:

[0877] During the interview, the interviewer asks the user questions through a neutral voice and virtual character, and the user answers. This repeated process is carried out in real time on the server and the terminal. The input is the interviewer's questions and the user's answers, and the output is real-time alternating communication between them.

[0878] Step 10:

[0879] After the interview is over, the user presses the "End" button to end the interview. This causes the server to stop processing the audio and video streams and end the session. The input is the end command, and the output is the stopped state of the audio and video streams.

[0880] (Application example 1)

[0881] 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."

[0882] Audio and video neutralization technology was invented to eliminate gender bias in online interview systems. Similarly, various voice operations and interactions are increasingly required in autonomous vehicles, where gender bias also presents a problem. Conventional in-vehicle interactive systems can contain unconscious gender bias in voice assistants and display displays, resulting in a lack of fairness for passengers and drivers.

[0883] 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.

[0884] In this invention, the server includes means for receiving a user's voice stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for outputting the processed voice stream and video stream to an interactive system of the autonomous vehicle, means for outputting the user's gender-neutral voice to an in-vehicle speaker, and means for displaying the gender-neutral avatar on an in-vehicle display using face detection, thereby eliminating gender bias in voice operations and display views in the autonomous vehicle and enabling fair and neutral interactions.

[0885] "User" refers to a person who uses a system or application.

[0886] An "audio stream" refers to a continuously transmitted stream of audio data.

[0887] "Gender-neutral voice" refers to a voice that has vocal characteristics that are not biased towards a specific gender.

[0888] A "video stream" refers to a continuously transmitted stream of video data.

[0889] "Face detection" refers to the process of recognizing the location and size of a face in a particular frame within a video stream.

[0890] A "gender-neutral avatar" refers to a virtual character whose visual elements are not gender-biased.

[0891] An "autonomous vehicle" refers to a vehicle that drives autonomously without the need for driver intervention.

[0892] An "interactive system" refers to a system that allows two-way communication with the user.

[0893] "In-vehicle speaker" refers to an audio output device installed inside a vehicle.

[0894] "Server" refers to a computer system that provides resources and processes data over a network.

[0895] "Sequential processing" refers to a method of processing data one by one in succession.

[0896] An "algorithm" refers to a method of solving a problem using specific computational procedures.

[0897] "Display" refers to a device or screen for displaying images or videos.

[0898] The present invention provides a technology for eliminating gender bias in interactive systems within autonomous vehicles. Specific embodiments of this system are described below.

[0899] Hardware and software used

[0900] 1. Hardware

[0901] Server: Located remotely and responsible for processing the audio and video streams.

[0902] Vehicle microphone: A device for capturing the user's voice stream.

[0903] Dash camera: A device for capturing a user's video stream.

[0904] Car speaker: A device for outputting converted neutral audio.

[0905] In-car display: A screen for displaying a gender-neutral avatar in the car.

[0906] 2. Software

[0907] Speech recognition engine (e.g., Google Speech-to-Text API, Amazon Transcribe): Recognizes the user's audio stream and converts it into text.

[0908] Voice conversion algorithms (e.g., Voice Conversion NN(VCNN)): convert a voice stream into a neutral voice.

[0909] Face detection libraries (e.g. OpenCV): Detect faces in the captured video stream.

[0910] Avatar generation engine (e.g. Unity): Generates gender-neutral avatars based on face detection results.

[0911] User interface frameworks (e.g., React Native, Flutter): for building the front-end of in-car interactive systems.

[0912] Server Procedure

[0913] 1. Audio capture and conversion

[0914] The server receives the audio stream transmitted from the vehicle microphone.

[0915] The received audio stream is converted to a gender-neutral voice using a voice conversion algorithm, which adjusts pitch and tone to eliminate gender bias.

[0916] The converted neutral voice is output to the car's speakers.

[0917] 2. Video Stream Processing

[0918] The server receives the video stream transmitted from the in-car camera.

[0919] Each frame in the video stream is processed in real time using a face detection library.

[0920] A gender-neutral avatar generated by an avatar generation engine is superimposed on the detected face.

[0921] The generated avatar is displayed on the in-car display.

[0922] Specific examples

[0923] Audio capture and conversion examples

[0924] When a user says "Start navigation" to the in-car assistant, the server converts this speech into a gender-neutral voice in real time. The in-car speaker plays "Start navigation" in a consistent, gender-neutral tone. This real-time conversion is achieved using a voice conversion algorithm.

[0925] Video stream processing example

[0926] When a user enters a vehicle, the dashcam starts capturing the video stream. The server detects faces in each frame and displays a neutral avatar. For example, if the user is smiling, the avatar will adjust to smile as well.

[0927] Example prompts

[0928] "Convert to gender-neutral voice. This voice is a command for your in-car assistant."

[0929] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0930] Step 1: Capture the audio stream

[0931] The user speaks a voice command to the in-car assistant.

[0932] Input: User's voice captured by the vehicle's microphone.

[0933] Processing: The vehicle microphone sends the audio stream to the server.

[0934] Output: Real-time audio stream sent to the server.

[0935] Step 2: Neutralizing the Audio Stream

[0936] The server inputs the received audio stream into a voice conversion algorithm.

[0937] Input: The audio stream sent to the server.

[0938] Processing: A voice conversion algorithm (e.g., Voice Conversion NN (VCNN)) adjusts the pitch and tone of the voice to convert it into a neutral voice.

[0939] Output: Neutral converted audio stream.

[0940] Step 3: Output audio

[0941] The server sends a neutral audio stream to the car's speakers.

[0942] Input: Neutral converted audio stream.

[0943] Processing: A neutral audio stream is sent to the car speakers.

[0944] Output: Neutral voice played through the car speakers.

[0945] Step 4: Capturing the video stream

[0946] While the user is inside the car, the on-board camera captures a video stream.

[0947] Input: Video stream captured by the dashcam.

[0948] Processing: The video stream is sent to the server.

[0949] Output: Real-time video stream sent to the server.

[0950] Step 5: Face detection

[0951] The server inputs the received video stream into a face detection algorithm.

[0952] Input: The video stream sent to the server.

[0953] Processing: A face detection algorithm (e.g. OpenCV) detects the position and size of the user's face in each frame.

[0954] Output: Face position and size information for each frame.

[0955] Step 6: Avatar generation

[0956] The server inputs the face detection information into an avatar generation engine to generate a gender-neutral avatar.

[0957] Input: Face position and size information.

[0958] Processing: An avatar generation engine (e.g., Unity) generates a gender-neutral avatar based on the detected facial features and overlays it on each frame.

[0959] Output: A frame with an androgynous avatar superimposed.

[0960] Step 7: Output the video stream

[0961] The server transmits the video stream with the avatar superimposed on it to an in-car display.

[0962] Input: A video frame with an androgynous avatar superimposed on it.

[0963] Processing: The video stream is sent to the in-car display.

[0964] Output: A video stream with a gender-neutral avatar displayed on the in-car display.

[0965] 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.

[0966] The present invention relates to an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[0967] Processing on the server

[0968] The server receives the voice stream sent from the user's device. The received voice stream is converted to a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the voice to create a gender-neutral voice. The converted voice stream is then sent to the interviewer's device.

[0969] Furthermore, the server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, and this recognized emotion data is also sent to the interviewer's device.

[0970] For example, if a user speaks in an excited tone, the emotion engine will recognize this as "excited" and convey that information to the interviewer.

[0971] Processing on the device

[0972] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[0973] Furthermore, the device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, and changes the avatar's facial expression based on the recognized emotions.

[0974] For example, if the user is smiling, the emotion engine will recognize this as "happiness" and adjust the avatar to smile as well.

[0975] User processing

[0976] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the system to capture the user's voice and video. The user's voice is converted to a neutral voice and a neutral avatar is displayed on the video. Furthermore, the system recognizes the user's emotions and relays this information to the interviewer in real time.

[0977] For example, when a user introduces themselves in an online interview, their voice is conveyed as a neutral voice and a neutral avatar is displayed on the video screen. If the user is nervous, the emotion engine detects this and tells the interviewer that they are "nervous."

[0978] Through the above processing, the present invention eliminates gender bias and reflects the user's feelings, thereby achieving fair and objective interview evaluation.

[0979] The processing flow will be explained below.

[0980] Processing on the server

[0981] Step 1:

[0982] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[0983] Step 2:

[0984] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[0985] Step 3:

[0986] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[0987] Step 4:

[0988] The server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, analyzing voice characteristics such as tempo and pitch to identify the type of emotion.

[0989] Step 5:

[0990] The server sends the recognized emotion data to the interviewer's device, and transfers the emotion data in real time as additional information.

[0991] Processing on the device

[0992] Step 1:

[0993] The device receives a video stream from the user's camera, using OpenCV and other libraries to capture video frames from the camera.

[0994] Step 2:

[0995] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[0996] Step 3:

[0997] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[0998] Step 4:

[0999] The device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, for example, by analyzing facial movements and changes in facial features to identify emotions.

[1000] Step 5:

[1001] The device will then change the avatar's facial expression based on the recognized emotion, for example, making the avatar smile if the user is smiling.

[1002] Step 6:

[1003] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[1004] User processing

[1005] Step 1:

[1006] The user logs into the online interview system, accessing the system using a designated application or web portal.

[1007] Step 2:

[1008] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[1009] Step 3:

[1010] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[1011] Step 4:

[1012] The user's emotions are recognized in real time and communicated to the interviewer. For example, if the user is nervous, that information is displayed to the interviewer.

[1013] The above processing steps eliminate gender bias and allow for an online interview that reflects emotions, enabling a fair and objective evaluation.

[1014] Example 2

[1015] 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."

[1016] In conventional online interview systems, gender bias is often present in the user's voice and video, making it difficult to conduct fair evaluations. Furthermore, there is a lack of a way to recognize the user's emotions in real time and convey them to the interviewer, which means that interview evaluations tend to be subjective. There is a need to solve these issues and achieve fair and objective interview evaluations.

[1017] 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.

[1018] In this invention, the server includes means for receiving a user's voice stream and converting it to a gender-neutral voice, means for analyzing the voice stream and recognizing emotions, and means for outputting the recognized emotion data. This enables the elimination of gender bias and real-time recognition of user emotions. The server also includes means for acquiring a user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for analyzing the video stream and recognizing emotions from the user's facial expressions, means for changing the avatar's facial expressions based on the recognized emotions, and means for outputting the processed voice stream and video stream. This enables fair and objective interview evaluation.

[1019] An "audio stream" is a data flow that transmits a user's voice as digital data in real time.

[1020] A "gender-neutral voice" is a voice that is not biased towards a particular gender and makes it difficult for the listener to identify the gender.

[1021] A "means for transforming" is an algorithm or device that changes input data into a different format or characteristic.

[1022] A "video stream" is a data flow that transmits a user's video as digital data in real time.

[1023] "Sequential processing" is a method of processing continuously input data one by one.

[1024] A "means for detecting faces" is an algorithm or device for identifying and locating human faces within video data.

[1025] An "avatar" is a virtual character or image that is displayed in a digital space on behalf of a user.

[1026] A "superimposing means" is an algorithm or device for displaying separate images or data in an overlapping manner.

[1027] Analytical methods are methods for examining data in detail and extracting specific information or patterns from it.

[1028] "Means for recognizing emotions" refers to algorithms or devices for identifying a user's emotional state from audio or video data.

[1029] "Emotional data" is digital information about a user's perceived emotional state.

[1030] A "means for changing facial expressions" is an algorithm or device for changing the facial expression of an avatar in response to the emotional state of the user.

[1031] "Output means" refers to a method or device for providing processed data to an external entity.

[1032] This invention is an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[1033] Processing on the server

[1034] The server receives the user's voice stream and converts it into a gender-neutral voice. First, it receives the voice stream sent from the user's device. This voice stream is converted into a gender-neutral voice in real time using a dedicated voice conversion algorithm. For voice conversion, a Python-based voice processing library (e.g., librosa) is used. The pitch and tone of the voice are adjusted to produce a voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's device.

[1035] The server then analyzes the audio stream and uses an emotion engine to recognize the user's emotions. The emotion recognition uses the Microsoft Azure Emotion API. The analyzed emotion data is sent to the interviewer's device in real time. For example, if the user speaks in an excited tone, the emotion engine recognizes this as "excitement," and the information is displayed on the interviewer's device.

[1036] Processing on the device

[1037] The device acquires the user's video stream. Each frame in the acquired video stream is processed sequentially in real time. The user's face is detected using a dedicated face detection algorithm. This process uses the OpenCV library. After face detection is complete, a neutral avatar image is superimposed on the detected face. Unity or Unreal Engine is used to generate and display the avatar.

[1038] The device also analyzes the video stream to recognize emotions from the user's facial expressions. Based on the recognized emotion, the avatar's facial expression changes. This emotion recognition again uses the Microsoft Azure Emotion API. For example, if the user is smiling, the emotion engine recognizes this as "joy," and the avatar is adjusted to smile as well.

[1039] User processing

[1040] The user logs into the online interview system and grants permission to use the camera and microphone. This permission allows the user's voice and video to be captured in real time. The user's voice is converted to a neutral voice, and a neutral avatar is displayed on the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the interviewer in real time.

[1041] As a concrete example, if a user enters the following prompt into an online interview system:

[1042] "Hello, I'm Taro Tanaka. I currently work as a data scientist, with particular expertise in machine learning and data analysis."

[1043] By entering this prompt, the system can accurately convey the user's gender-neutral voice and avatar to the interviewer.

[1044] From the above, it can be seen that this invention is a system that eliminates gender bias and reflects the user's emotions, thereby realizing fair and objective interview evaluation.

[1045] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1046] Step 1:

[1047] The server receives the audio stream sent from the user's device. Specifically, the audio data spoken by the user is sent to the server in digital form. This received data is input into the server's audio processing module. The output is audio data for conversion processing. For example, if the user says "hello," the audio data reaches the server.

[1048] Step 2:

[1049] The server converts the received audio stream into a gender-neutral voice. Here, it uses a Python-based audio processing library (e.g., librosa). It adjusts the pitch and tone of the input audio data to generate a gender-neutral voice that is not biased toward a specific gender. This converted audio data is then passed on to the next process. For example, a high-pitched male voice or a low-pitched female voice can be converted into a neutral voice.

[1050] Step 3:

[1051] The server sends the converted gender-neutral voice stream to the interviewer's device. The input is gender-neutral voice data, and the output is voice data sent to the interviewer's device. This allows the interviewer to receive voice without gender bias.

[1052] Step 4:

[1053] The server analyzes the audio stream and recognizes the user's emotion using an emotion engine (e.g., Microsoft Azure Emotion API). The input is the user's voice data, and the output is the recognized emotion data. For example, if the user speaks in an excited tone, the emotion engine will recognize this as "excited."

[1054] Step 5:

[1055] The server sends the recognized emotion data to the interviewer's device. The input is the recognized emotion data, and the output is the emotion information sent to the interviewer's device. This allows the interviewer to grasp the user's emotional state in real time.

[1056] Step 6:

[1057] The device acquires a video stream from the user's camera. The input is video data from the camera, and the output is video stream data. For example, when a user introduces themselves to the camera, the video is captured.

[1058] Step 7:

[1059] The device processes each frame in the captured video stream in real time and detects the user's face using a dedicated face detection algorithm (e.g., OpenCV). The input is the video stream data, and the output is the face coordinate data.

[1060] Step 8:

[1061] The device superimposes a neutral avatar image on the detected face. The input is face coordinate data and avatar image data, and the output is a video frame showing the avatar. This process is performed using Unity or Unreal Engine.

[1062] Step 9:

[1063] The device analyzes the video stream and recognizes emotions from the user's facial expressions. Emotion recognition uses the Microsoft Azure Emotion API. The input is the user's video stream data, and the output is the recognized emotion data. For example, if the user is smiling, the emotion engine recognizes this as "joy."

[1064] Step 10:

[1065] The device changes the facial expression of the avatar based on the recognized emotion. The input is the recognized emotion data, and the output is a video frame of the avatar with the changed facial expression. This allows the user's emotion to be visually conveyed to the interviewer.

[1066] Step 11:

[1067] A user logs in to the online interview system and grants permission to use the camera and microphone. The input is the user's login information, and the output is the access permission to the system. This allows the user's voice and video to be captured by the system.

[1068] Step 12:

[1069] The user's voice and video are captured in real time. The input is data from the camera and microphone, and the output is an audio stream and a video stream. For example, when a user introduces themselves, their voice and video are captured by the system.

[1070] Step 13:

[1071] The system converts the user's voice into a gender-neutral voice and displays a gender-neutral avatar on the screen. The input is the user's voice and video data, and the output is a gender-neutral voice and video data with an avatar.

[1072] Step 14:

[1073] The system recognizes the user's emotions in real time and conveys that information to the interviewer. The input is audio and video stream data, and the output is emotion data sent to the interviewer's device. For example, if the user is nervous, the emotion engine recognizes this as "nervous" and conveys that information to the interviewer.

[1074] (Application example 2)

[1075] 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."

[1076] Current online interview systems and other remote communication systems can be affected by gender bias. Furthermore, because users' emotions are not clearly conveyed, unfair evaluations may be made during interviews or customer service. It is necessary to solve these problems and achieve fair and objective communication.

[1077] 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 receiving a user's audio stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for recognizing emotions from the user's audio and video streams, means for changing the avatar's facial expression based on the recognized emotions, and means for outputting the processed audio and video streams. This eliminates gender bias and enables objective evaluation and fair communication that accurately reflects the user's emotions.

[1078] An "audio stream" is data that is continuously transmitted by converting the user's voice into a digital signal.

[1079] "Neutral voice" refers to a voice whose pitch and tone are adjusted to create a neutral voice that is not biased towards a specific gender.

[1080] A "video stream" is continuous video data captured from a user's camera.

[1081] "Sequential processing" refers to processing acquired data continuously in real time.

[1082] "Face detection" is the process of analyzing each frame in a video stream and identifying human faces.

[1083] An "avatar" is a virtual image of a person that is displayed on the screen in place of the user.

[1084] "Emotion recognition" refers to the process of analyzing and determining a user's emotional state from their audio and video streams.

[1085] "Changing facial expression" means dynamically changing the avatar's facial expression according to the recognized emotion.

[1086] "Output" refers to the action of displaying or transmitting processed data on another terminal or screen.

[1087] A "system" is a set of devices or programs in which multiple means operate in cooperation with each other.

[1088] The present invention provides a fair customer service experience in a virtual store based on an online interview system with gender-neutral voice conversion and emotion recognition functions. In this embodiment, a user, a terminal, and a server work together.

[1089] Processing on the server

[1090] The server receives the user's voice stream and converts it into a gender-neutral voice. This voice is generated by applying a voice conversion algorithm to ensure a gender-neutral voice. The server then uses an emotion engine to recognize emotions from the user's voice stream. This emotion data plays an important role in the customer service system of the virtual store.

[1091] Processing on the device

[1092] The device acquires a video stream from the user's camera and processes each frame in this video stream in real time. It uses a face detection algorithm such as OpenCV to detect the user's face in each frame. It then superimposes a neutral avatar image on the detected face. The device then analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions. It also has the ability to change the avatar's facial expression based on the recognized emotion.

[1093] User processing

[1094] The user logs into the virtual store customer service system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed in the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the store staff in real time. For example, if the user is excited about a product, the customer service system can use this information to make appropriate product introductions and suggestions, thereby increasing the user's desire to purchase.

[1095] Hardware and Software

[1096] Hardware: Camera, microphone, PC

[1097] Software: OpenCV, PyAudio, Emotion Recognizer (our own emotion recognition engine), Gender Neutral Voice (our own voice conversion module)

[1098] Specific examples

[1099] When a customer asks, "Tell me about this product," the system responds in a neutral voice, saying, "This product uses the latest technology and features X features." At the same time, if the customer is excited, the system notifies the store staff, who can then provide more detailed information or special offers.

[1100] Prompt Sentence Examples

[1101] "Give us an example of a process that detects faces in every video frame, overlays them with an androgynous avatar, and performs customer emotion recognition."

[1102] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1103] Step 1:

[1104] The user logs in to the virtual store customer service system and grants permission to use the camera and microphone.

[1105] Input: User's camera image and microphone audio

[1106] Action: Allow access to camera and microphone devices

[1107] Output: You can get the user's audio and video streams.

[1108] Step 2:

[1109] The device acquires the video stream from the camera and starts processing it sequentially in real time.

[1110] Input: Video stream from camera

[1111] Processing: Capture video stream frame by frame

[1112] Output: Individual video frames

[1113] Step 3:

[1114] The device uses OpenCV to detect the user's face in each video frame.

[1115] Input: Individual video frames

[1116] Processing: Apply OpenCV face detection algorithm to identify face area

[1117] Output: Video frames where faces are detected and face coordinate information

[1118] Step 4:

[1119] The device overlays a neutral avatar image onto the detected face.

[1120] Input: Coordinate information of face area

[1121] Processing: Overlay the avatar image onto the detected face area

[1122] Output: Video frame with avatar superimposed

[1123] Step 5:

[1124] The user's voice stream is also simultaneously captured by the terminal and sent to the server.

[1125] Input: Audio stream from microphone

[1126] Processing: Capture the audio stream and send it to the server

[1127] Output: The audio stream sent to the server

[1128] Step 6:

[1129] The server receives the audio stream and converts it to a neutral voice.

[1130] Input: The user's audio stream

[1131] Processing: Applying voice conversion algorithms to generate a neutral voice

[1132] Output: Neutral audio stream

[1133] Step 7:

[1134] The server utilizes an emotion engine to recognize the user's emotion from the audio stream.

[1135] Input: Neutral audio stream

[1136] Processing: Applying emotion engine to analyze emotional state from voice data

[1137] Output: Recognized emotion data

[1138] Step 8:

[1139] The recognized emotion data is transmitted to the terminal.

[1140] Input: Recognized emotion data

[1141] Processing: Emotion data is transferred to the device

[1142] Output: Emotion data sent to the device

[1143] Step 9:

[1144] The device analyzes the video stream and recognizes emotions from the user's facial expressions.

[1145] Input: Video frame

[1146] Processing: Apply video stream analysis and emotion engine to recognize emotions from facial expression data

[1147] Output: Emotion data from recognized facial expressions

[1148] Step 10:

[1149] The device changes the avatar's facial expression based on the recognized emotion.

[1150] Input: Emotion data recognized from voice and facial expressions

[1151] Processing: Dynamically change avatar facial expressions

[1152] Output: Video frames with an avatar superimposed with facial expressions corresponding to the emotion.

[1153] Step 11:

[1154] The processed audio and video streams are output in real time to shop staff in the virtual store.

[1155] Input: Neutral audio stream and video frames with avatars superimposed

[1156] Processing: Transfer of audio and video streams

[1157] Output: Audio and video streams displayed on shop staff's terminals

[1158] 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.

[1159] 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.

[1160] 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.

[1161] [Fourth embodiment]

[1162] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1163] 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.

[1164] 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).

[1165] 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.

[1166] 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.

[1167] 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).

[1168] 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.

[1169] 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.

[1170] 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.

[1171] 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.

[1172] 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.

[1173] 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.

[1174] 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."

[1175] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[1176] server

[1177] The server has the function of receiving the voice stream sent by the user. The received voice stream is converted into a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the input voice to generate a gender-neutral voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's terminal.

[1178] For example, if a user speaks in a high-pitched voice, the voice will be changed by the server to an appropriate pitch and transmitted to the interviewer in a neutral tone.

[1179] Terminal

[1180] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[1181] Specifically, a face detection algorithm identifies the position and size of the face within the frame, then resizes the avatar image to fit that size. The resized avatar is then overlaid on top of the original face image. This modified frame is displayed to both the user and the interviewer.

[1182] For example, if the user is smiling in the displayed frame, the avatar will be adjusted to appear smiling as well.

[1183] User

[1184] Users log in to the online interview system and grant permission to use the camera and microphone, which allows the system to capture their voice and video. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed on the video. This allows users to be interviewed without any bias regarding their gender.

[1185] For example, when a user introduces themselves in an online interview, their voice will be transmitted as a gender-neutral voice and a gender-neutral avatar will be displayed on the video screen. The interviewer will see and hear this gender-neutral voice and avatar, allowing the interview to proceed without being influenced by gender bias.

[1186] This invention allows interviewers to evaluate candidates without gender bias, thereby realizing fair and objective interviews.

[1187] The processing flow will be explained below.

[1188] Processing on the server

[1189] Step 1:

[1190] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[1191] Step 2:

[1192] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[1193] Step 3:

[1194] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[1195] Processing on the device

[1196] Step 1:

[1197] The device receives a video stream from the user's camera and uses OpenCV and other libraries to capture video frames from the camera.

[1198] Step 2:

[1199] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[1200] Step 3:

[1201] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[1202] Step 4:

[1203] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[1204] User processing

[1205] Step 1:

[1206] The user logs into the online interview system, accessing the system using a designated application or web portal.

[1207] Step 2:

[1208] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[1209] Step 3:

[1210] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[1211] The above processing steps enable online interviews that are free of gender bias.

[1212] Example 1

[1213] 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."

[1214] In conventional online interview systems, gender bias can affect evaluations. A particular problem is that interviewers may unconsciously have preconceived notions based on the candidate's gender. This makes it difficult to conduct fair and objective evaluations, and it is difficult to accurately determine the candidate's true abilities and aptitude.

[1215] 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.

[1216] In this invention, the server includes means for receiving the user's voice data and converting it into a gender-neutral voice, means for acquiring the user's video data and sequentially processing it to detect the user's face, and means for superimposing a gender-neutral virtual character on the detected face, thereby eliminating gender bias and enabling fair and objective interviews.

[1217] A "user" is a person who uses the online interview system, accesses the system, and provides their own audio and video.

[1218] "Audio Data" refers to audio information collected through the user's microphone, which is sent to the server for processing.

[1219] "Gender-neutral voice" refers to voices with pitch and tone adjusted so that they are not associated with a specific gender.

[1220] "Video data" refers to video information collected through a user's camera, material that is processed frame by frame.

[1221] "Sequential processing" refers to a method of processing input data sequentially in real time.

[1222] "Face detection" refers to the process of identifying the location and size of a user's face from frames within video data.

[1223] "Virtual characters" are neutral images or animations that are superimposed on the user's face and are designed to avoid any gender-related associations.

[1224] "Superimposition" is the process of superimposing one image or object on top of another.

[1225] "Gender bias" is a term that refers to preconceived notions or prejudices against a particular gender, and is an unfair factor that influences evaluations and judgments.

[1226] The present invention relates to an online interview system for eliminating gender bias. A specific embodiment of this system will be described below.

[1227] server

[1228] The server has hardware and software to receive the voice data sent by the user. After receiving the voice data, the server uses the Vocoder algorithm to adjust the pitch and tone of the voice and convert it into a neutral voice. This converted voice data is then sent to the interviewer's device in real time.

[1229] For example, if a user provides a normal speaking voice, the voice is analyzed by the server and normalized to a standard pitch and tone.

[1230] Terminal

[1231] The device is equipped with hardware and software to capture video data from the user's camera. The captured video data is then processed sequentially using a face detection algorithm, such as OpenCV or Dlib, to identify the location and size of the user's face and overlay a gender-neutral virtual character on it.

[1232] This superimposed video data is displayed in real time to both the user and the interviewer. For example, when the user speaks with a smile, the virtual character is adjusted so that it also appears smiling based on that facial expression.

[1233] User

[1234] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured and processed on the server and device. The user's voice is converted to a gender-neutral voice, and a gender-neutral virtual character is superimposed on the video. This system allows the user to be interviewed without any gender bias.

[1235] For example, when a user introduces themselves in an online interview, their voice is converted to a gender-neutral voice on the server side, and a gender-neutral virtual character is displayed on the video screen. Through this gender-neutral voice and virtual character, the interviewer can conduct the interview without being influenced by gender bias.

[1236] Examples of prompt statements

[1237] Explain an online interview system to eliminate gender bias. Describe in detail the specific operation and examples of each component (server, terminal, user).

[1238] This invention is expected to enable fair and objective interviews and improve the accuracy of evaluations.

[1239] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1240] Step 1:

[1241] The user logs in to the online interview system. They enter their user ID and password on the login screen to access the system. The system then authenticates the user and sets up the system to start the interview session. The input is the user's authentication information, and the output is the system access rights after successful login.

[1242] Step 2:

[1243] The user grants permission to access the camera and microphone. The system displays a pop-up message, and the user must click "Allow" to access the device. The input is the user's permission, and the output is the connection status of the camera and microphone.

[1244] Step 3:

[1245] The user's device acquires video data from the camera and audio data from the microphone. These data are sent to the server as input data. For example, the camera captures video at 30 frames per second, and the microphone captures audio in real time. The input is the video and audio data, and the output is the data sent to the server.

[1246] Step 4:

[1247] The server receives the transmitted voice data and converts it into a gender-neutral voice using the Vocoder algorithm, adjusting the pitch and tone of the voice in real time to eliminate gender cues. The input is the original voice data, and the output is the converted gender-neutral voice data.

[1248] Step 5:

[1249] The converted neutral voice data is sent from the server to the interviewer's device in real time. The input is neutral voice data, and the output is the data sent to the interviewer's device. For example, if the user speaks in a high-pitched voice, the voice will be converted to a flat, neutral voice and delivered to the interviewer.

[1250] Step 6:

[1251] The device processes the video data acquired from the camera in real time and applies a face detection algorithm (e.g., OpenCV or Dlib) to identify the position and size of the user's face. The input is the original video data, and the output is the position and size information of the face.

[1252] Step 7:

[1253] Based on the detected face position and size, an image processing algorithm generates a gender-neutral virtual character and superimposes it on the original face image. The input is face position and size information, and the output is video data with the virtual character superimposed. For example, if a user is smiling while talking, the gender-neutral virtual character will also be displayed smiling.

[1254] Step 8:

[1255] The processed video data is displayed in real time to both the user and the interviewer. The input is video data with a virtual character superimposed, and the output is a video feed displayed on the user's and interviewer's devices.

[1256] Step 9:

[1257] During the interview, the interviewer asks the user questions through a neutral voice and virtual character, and the user answers. This repeated process is carried out in real time on the server and the terminal. The input is the interviewer's questions and the user's answers, and the output is real-time alternating communication between them.

[1258] Step 10:

[1259] After the interview is over, the user presses the "End" button to end the interview. This causes the server to stop processing the audio and video streams and end the session. The input is the end command, and the output is the stopped state of the audio and video streams.

[1260] (Application example 1)

[1261] 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."

[1262] Audio and video neutralization technology was invented to eliminate gender bias in online interview systems. Similarly, various voice operations and interactions are increasingly required in autonomous vehicles, where gender bias also presents a problem. Conventional in-vehicle interactive systems can contain unconscious gender bias in voice assistants and display displays, resulting in a lack of fairness for passengers and drivers.

[1263] 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.

[1264] In this invention, the server includes means for receiving a user's voice stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for outputting the processed voice stream and video stream to an interactive system of the autonomous vehicle, means for outputting the user's gender-neutral voice to an in-vehicle speaker, and means for displaying the gender-neutral avatar on an in-vehicle display using face detection, thereby eliminating gender bias in voice operations and display views in the autonomous vehicle and enabling fair and neutral interactions.

[1265] "User" refers to a person who uses a system or application.

[1266] An "audio stream" refers to a continuously transmitted stream of audio data.

[1267] "Gender-neutral voice" refers to a voice that has vocal characteristics that are not biased towards a specific gender.

[1268] A "video stream" refers to a continuously transmitted stream of video data.

[1269] "Face detection" refers to the process of recognizing the location and size of a face in a particular frame within a video stream.

[1270] A "gender-neutral avatar" refers to a virtual character whose visual elements are not gender-biased.

[1271] An "autonomous vehicle" refers to a vehicle that drives autonomously without the need for driver intervention.

[1272] An "interactive system" refers to a system that allows two-way communication with the user.

[1273] "In-vehicle speaker" refers to an audio output device installed inside a vehicle.

[1274] "Server" refers to a computer system that provides resources and processes data over a network.

[1275] "Sequential processing" refers to a method of processing data one by one in succession.

[1276] An "algorithm" refers to a method of solving a problem using specific computational procedures.

[1277] "Display" refers to a device or screen for displaying images or videos.

[1278] The present invention provides a technology for eliminating gender bias in interactive systems within autonomous vehicles. Specific embodiments of this system are described below.

[1279] Hardware and software used

[1280] 1. Hardware

[1281] Server: Located remotely and responsible for processing the audio and video streams.

[1282] Vehicle microphone: A device for capturing the user's voice stream.

[1283] Dash camera: A device for capturing a user's video stream.

[1284] Car speaker: A device for outputting converted neutral audio.

[1285] In-car display: A screen for displaying a gender-neutral avatar in the car.

[1286] 2. Software

[1287] Speech recognition engine (e.g., Google Speech-to-Text API, Amazon Transcribe): Recognizes the user's audio stream and converts it into text.

[1288] Voice conversion algorithms (e.g., Voice Conversion NN(VCNN)): convert a voice stream into a neutral voice.

[1289] Face detection libraries (e.g. OpenCV): Detect faces in the captured video stream.

[1290] Avatar generation engine (e.g. Unity): Generates gender-neutral avatars based on face detection results.

[1291] User interface frameworks (e.g., React Native, Flutter): for building the front-end of in-car interactive systems.

[1292] Server Procedure

[1293] 1. Audio capture and conversion

[1294] The server receives the audio stream transmitted from the vehicle microphone.

[1295] The received audio stream is converted to a gender-neutral voice using a voice conversion algorithm, which adjusts pitch and tone to eliminate gender bias.

[1296] The converted neutral voice is output to the car's speakers.

[1297] 2. Video Stream Processing

[1298] The server receives the video stream transmitted from the in-car camera.

[1299] Each frame in the video stream is processed in real time using a face detection library.

[1300] A gender-neutral avatar generated by an avatar generation engine is superimposed on the detected face.

[1301] The generated avatar is displayed on the in-car display.

[1302] Specific examples

[1303] Audio capture and conversion examples

[1304] When a user says "Start navigation" to the in-car assistant, the server converts this speech into a gender-neutral voice in real time. The in-car speaker plays "Start navigation" in a consistent, gender-neutral tone. This real-time conversion is achieved using a voice conversion algorithm.

[1305] Video stream processing example

[1306] When a user enters a vehicle, the dashcam starts capturing the video stream. The server detects faces in each frame and displays a neutral avatar. For example, if the user is smiling, the avatar will adjust to smile as well.

[1307] Example prompts

[1308] "Convert to gender-neutral voice. This voice is a command for your in-car assistant."

[1309] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1310] Step 1: Capture the audio stream

[1311] The user speaks a voice command to the in-car assistant.

[1312] Input: User's voice captured by the vehicle's microphone.

[1313] Processing: The vehicle microphone sends the audio stream to the server.

[1314] Output: Real-time audio stream sent to the server.

[1315] Step 2: Neutralizing the Audio Stream

[1316] The server inputs the received audio stream into a voice conversion algorithm.

[1317] Input: The audio stream sent to the server.

[1318] Processing: A voice conversion algorithm (e.g., Voice Conversion NN (VCNN)) adjusts the pitch and tone of the voice to convert it into a neutral voice.

[1319] Output: Neutral converted audio stream.

[1320] Step 3: Output audio

[1321] The server sends a neutral audio stream to the car's speakers.

[1322] Input: Neutral converted audio stream.

[1323] Processing: A neutral audio stream is sent to the car speakers.

[1324] Output: Neutral voice played through the car speakers.

[1325] Step 4: Capturing the video stream

[1326] While the user is inside the car, the on-board camera captures a video stream.

[1327] Input: Video stream captured by the dashcam.

[1328] Processing: The video stream is sent to the server.

[1329] Output: Real-time video stream sent to the server.

[1330] Step 5: Face detection

[1331] The server inputs the received video stream into a face detection algorithm.

[1332] Input: The video stream sent to the server.

[1333] Processing: A face detection algorithm (e.g. OpenCV) detects the position and size of the user's face in each frame.

[1334] Output: Face position and size information for each frame.

[1335] Step 6: Avatar generation

[1336] The server inputs the face detection information into an avatar generation engine to generate a gender-neutral avatar.

[1337] Input: Face position and size information.

[1338] Processing: An avatar generation engine (e.g., Unity) generates a gender-neutral avatar based on the detected facial features and overlays it on each frame.

[1339] Output: A frame with an androgynous avatar superimposed.

[1340] Step 7: Output the video stream

[1341] The server transmits the video stream with the avatar superimposed on it to an in-car display.

[1342] Input: A video frame with an androgynous avatar superimposed on it.

[1343] Processing: The video stream is sent to the in-car display.

[1344] Output: A video stream with a gender-neutral avatar displayed on the in-car display.

[1345] 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.

[1346] The present invention relates to an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[1347] Processing on the server

[1348] The server receives the voice stream sent from the user's device. The received voice stream is converted to a gender-neutral voice using a dedicated voice conversion algorithm. This voice conversion algorithm adjusts the pitch and tone of the voice to create a gender-neutral voice. The converted voice stream is then sent to the interviewer's device.

[1349] Furthermore, the server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, and this recognized emotion data is also sent to the interviewer's device.

[1350] For example, if a user speaks in an excited tone, the emotion engine will recognize this as "excited" and convey that information to the interviewer.

[1351] Processing on the device

[1352] The device captures a video stream from the user's camera. Each frame in the video stream is processed sequentially in real time. A dedicated face detection algorithm is used to detect the user's face in each frame. A neutral avatar image is then superimposed on the detected face.

[1353] Furthermore, the device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, and changes the avatar's facial expression based on the recognized emotions.

[1354] For example, if the user is smiling, the emotion engine will recognize this as "happiness" and adjust the avatar to smile as well.

[1355] User processing

[1356] The user logs into the online interview system and grants permission to use the camera and microphone. This allows the system to capture the user's voice and video. The user's voice is converted to a neutral voice and a neutral avatar is displayed on the video. Furthermore, the system recognizes the user's emotions and relays this information to the interviewer in real time.

[1357] For example, when a user introduces themselves in an online interview, their voice is conveyed as a neutral voice and a neutral avatar is displayed on the video screen. If the user is nervous, the emotion engine detects this and tells the interviewer that they are "nervous."

[1358] Through the above processing, the present invention eliminates gender bias and reflects the user's feelings, thereby achieving fair and objective interview evaluation.

[1359] The processing flow will be explained below.

[1360] Processing on the server

[1361] Step 1:

[1362] The server receives the audio stream sent from the user's device and acquires the audio data in real time via the network.

[1363] Step 2:

[1364] The server then feeds the received audio stream into a specific algorithm that converts it into a neutral voice, which, for example, adjusts the pitch of the voice and changes the tone.

[1365] Step 3:

[1366] The server then sends the converted neutral audio stream to the interviewer's device, transferring the data in real time over the network.

[1367] Step 4:

[1368] The server analyzes the user's voice stream and uses an emotion engine to recognize the user's emotions, analyzing voice characteristics such as tempo and pitch to identify the type of emotion.

[1369] Step 5:

[1370] The server sends the recognized emotion data to the interviewer's device, and transfers the emotion data in real time as additional information.

[1371] Processing on the device

[1372] Step 1:

[1373] The device receives a video stream from the user's camera, using OpenCV and other libraries to capture video frames from the camera.

[1374] Step 2:

[1375] The device processes the captured video frames sequentially and applies a face detection algorithm to identify the user's face within the frames.

[1376] Step 3:

[1377] The device resizes the neutral avatar image based on the position and size of the detected face, and then overlays the resized avatar image onto the face in the frame.

[1378] Step 4:

[1379] The device analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions, for example, by analyzing facial movements and changes in facial features to identify emotions.

[1380] Step 5:

[1381] The device will then change the avatar's facial expression based on the recognized emotion, for example, making the avatar smile if the user is smiling.

[1382] Step 6:

[1383] The device then displays the processed video frames on the user's and interviewer's displays, continuously processing the video in real time.

[1384] User processing

[1385] Step 1:

[1386] The user logs into the online interview system, accessing the system using a designated application or web portal.

[1387] Step 2:

[1388] The user allows camera and microphone access, and if permission is granted, the user's audio and video streams are captured.

[1389] Step 3:

[1390] The user answers questions from the interviewer during the online interview, and the captured voice is converted to a gender-neutral voice on the server, and a gender-neutral avatar is displayed on the video.

[1391] Step 4:

[1392] The user's emotions are recognized in real time and communicated to the interviewer. For example, if the user is nervous, that information is displayed to the interviewer.

[1393] The above processing steps eliminate gender bias and allow for an online interview that reflects emotions, enabling a fair and objective evaluation.

[1394] Example 2

[1395] 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."

[1396] In conventional online interview systems, gender bias is often present in the user's voice and video, making it difficult to conduct fair evaluations. Furthermore, there is a lack of a way to recognize the user's emotions in real time and convey them to the interviewer, which means that interview evaluations tend to be subjective. There is a need to solve these issues and achieve fair and objective interview evaluations.

[1397] 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.

[1398] In this invention, the server includes means for receiving a user's voice stream and converting it to a gender-neutral voice, means for analyzing the voice stream and recognizing emotions, and means for outputting the recognized emotion data. This enables the elimination of gender bias and real-time recognition of user emotions. The server also includes means for acquiring a user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for analyzing the video stream and recognizing emotions from the user's facial expressions, means for changing the avatar's facial expressions based on the recognized emotions, and means for outputting the processed voice stream and video stream. This enables fair and objective interview evaluation.

[1399] An "audio stream" is a data flow that transmits a user's voice as digital data in real time.

[1400] A "gender-neutral voice" is a voice that is not biased towards a particular gender and makes it difficult for the listener to identify the gender.

[1401] A "means for transforming" is an algorithm or device that changes input data into a different format or characteristic.

[1402] A "video stream" is a data flow that transmits a user's video as digital data in real time.

[1403] "Sequential processing" is a method of processing continuously input data one by one.

[1404] A "means for detecting faces" is an algorithm or device for identifying and locating human faces within video data.

[1405] An "avatar" is a virtual character or image that is displayed in a digital space on behalf of a user.

[1406] A "superimposing means" is an algorithm or device for displaying separate images or data in an overlapping manner.

[1407] Analytical methods are methods for examining data in detail and extracting specific information or patterns from it.

[1408] "Means for recognizing emotions" refers to algorithms or devices for identifying a user's emotional state from audio or video data.

[1409] "Emotional data" is digital information about a user's perceived emotional state.

[1410] A "means for changing facial expressions" is an algorithm or device for changing the facial expression of an avatar in response to the emotional state of the user.

[1411] "Output means" refers to a method or device for providing processed data to an external entity.

[1412] This invention is an online interview system that eliminates gender bias and recognizes and reflects the user's emotions. A specific embodiment of this system is described below.

[1413] Processing on the server

[1414] The server receives the user's voice stream and converts it into a gender-neutral voice. First, it receives the voice stream sent from the user's device. This voice stream is converted into a gender-neutral voice in real time using a dedicated voice conversion algorithm. For voice conversion, a Python-based voice processing library (e.g., librosa) is used. The pitch and tone of the voice are adjusted to produce a voice that is not biased towards a specific gender. The converted voice stream is then sent to the interviewer's device.

[1415] The server then analyzes the audio stream and uses an emotion engine to recognize the user's emotions. The emotion recognition uses the Microsoft Azure Emotion API. The analyzed emotion data is sent to the interviewer's device in real time. For example, if the user speaks in an excited tone, the emotion engine recognizes this as "excitement," and the information is displayed on the interviewer's device.

[1416] Processing on the device

[1417] The device acquires the user's video stream. Each frame in the acquired video stream is processed sequentially in real time. The user's face is detected using a dedicated face detection algorithm. This process uses the OpenCV library. After face detection is complete, a neutral avatar image is superimposed on the detected face. Unity or Unreal Engine is used to generate and display the avatar.

[1418] The device also analyzes the video stream to recognize emotions from the user's facial expressions. Based on the recognized emotion, the avatar's facial expression changes. This emotion recognition again uses the Microsoft Azure Emotion API. For example, if the user is smiling, the emotion engine recognizes this as "joy," and the avatar is adjusted to smile as well.

[1419] User processing

[1420] The user logs into the online interview system and grants permission to use the camera and microphone. This permission allows the user's voice and video to be captured in real time. The user's voice is converted to a neutral voice, and a neutral avatar is displayed on the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the interviewer in real time.

[1421] As a concrete example, if a user enters the following prompt into an online interview system:

[1422] "Hello, I'm Taro Tanaka. I currently work as a data scientist, with particular expertise in machine learning and data analysis."

[1423] By entering this prompt, the system can accurately convey the user's gender-neutral voice and avatar to the interviewer.

[1424] From the above, it can be seen that this invention is a system that eliminates gender bias and reflects the user's emotions, thereby realizing fair and objective interview evaluation.

[1425] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1426] Step 1:

[1427] The server receives the audio stream sent from the user's device. Specifically, the audio data spoken by the user is sent to the server in digital form. This received data is input into the server's audio processing module. The output is audio data for conversion processing. For example, if the user says "hello," the audio data reaches the server.

[1428] Step 2:

[1429] The server converts the received audio stream into a gender-neutral voice. Here, it uses a Python-based audio processing library (e.g., librosa). It adjusts the pitch and tone of the input audio data to generate a gender-neutral voice that is not biased toward a specific gender. This converted audio data is then passed on to the next process. For example, a high-pitched male voice or a low-pitched female voice can be converted into a neutral voice.

[1430] Step 3:

[1431] The server sends the converted gender-neutral voice stream to the interviewer's device. The input is gender-neutral voice data, and the output is voice data sent to the interviewer's device. This allows the interviewer to receive voice without gender bias.

[1432] Step 4:

[1433] The server analyzes the audio stream and recognizes the user's emotion using an emotion engine (e.g., Microsoft Azure Emotion API). The input is the user's voice data, and the output is the recognized emotion data. For example, if the user speaks in an excited tone, the emotion engine will recognize this as "excited."

[1434] Step 5:

[1435] The server sends the recognized emotion data to the interviewer's device. The input is the recognized emotion data, and the output is the emotion information sent to the interviewer's device. This allows the interviewer to grasp the user's emotional state in real time.

[1436] Step 6:

[1437] The device acquires a video stream from the user's camera. The input is video data from the camera, and the output is video stream data. For example, when a user introduces themselves to the camera, the video is captured.

[1438] Step 7:

[1439] The device processes each frame in the captured video stream in real time and detects the user's face using a dedicated face detection algorithm (e.g., OpenCV). The input is the video stream data, and the output is the face coordinate data.

[1440] Step 8:

[1441] The device superimposes a neutral avatar image on the detected face. The input is face coordinate data and avatar image data, and the output is a video frame showing the avatar. This process is performed using Unity or Unreal Engine.

[1442] Step 9:

[1443] The device analyzes the video stream and recognizes emotions from the user's facial expressions. Emotion recognition uses the Microsoft Azure Emotion API. The input is the user's video stream data, and the output is the recognized emotion data. For example, if the user is smiling, the emotion engine recognizes this as "joy."

[1444] Step 10:

[1445] The device changes the facial expression of the avatar based on the recognized emotion. The input is the recognized emotion data, and the output is a video frame of the avatar with the changed facial expression. This allows the user's emotion to be visually conveyed to the interviewer.

[1446] Step 11:

[1447] A user logs in to the online interview system and grants permission to use the camera and microphone. The input is the user's login information, and the output is the access permission to the system. This allows the user's voice and video to be captured by the system.

[1448] Step 12:

[1449] The user's voice and video are captured in real time. The input is data from the camera and microphone, and the output is an audio stream and a video stream. For example, when a user introduces themselves, their voice and video are captured by the system.

[1450] Step 13:

[1451] The system converts the user's voice into a gender-neutral voice and displays a gender-neutral avatar on the screen. The input is the user's voice and video data, and the output is a gender-neutral voice and video data with an avatar.

[1452] Step 14:

[1453] The system recognizes the user's emotions in real time and conveys that information to the interviewer. The input is audio and video stream data, and the output is emotion data sent to the interviewer's device. For example, if the user is nervous, the emotion engine recognizes this as "nervous" and conveys that information to the interviewer.

[1454] (Application example 2)

[1455] 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."

[1456] Current online interview systems and other remote communication systems can be affected by gender bias. Furthermore, because users' emotions are not clearly conveyed, unfair evaluations may be made during interviews or customer service. It is necessary to solve these problems and achieve fair and objective communication.

[1457] 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 receiving a user's audio stream and converting it into a gender-neutral voice, means for acquiring the user's video stream and sequentially processing it to detect the user's face, means for superimposing a gender-neutral avatar on the detected face, means for recognizing emotions from the user's audio and video streams, means for changing the avatar's facial expression based on the recognized emotions, and means for outputting the processed audio and video streams. This eliminates gender bias and enables objective evaluation and fair communication that accurately reflects the user's emotions.

[1458] An "audio stream" is data that is continuously transmitted by converting the user's voice into a digital signal.

[1459] "Neutral voice" refers to a voice whose pitch and tone are adjusted to create a neutral voice that is not biased towards a specific gender.

[1460] A "video stream" is continuous video data captured from a user's camera.

[1461] "Sequential processing" refers to processing acquired data continuously in real time.

[1462] "Face detection" is the process of analyzing each frame in a video stream and identifying human faces.

[1463] An "avatar" is a virtual image of a person that is displayed on the screen in place of the user.

[1464] "Emotion recognition" refers to the process of analyzing and determining a user's emotional state from their audio and video streams.

[1465] "Changing facial expression" means dynamically changing the avatar's facial expression according to the recognized emotion.

[1466] "Output" refers to the action of displaying or transmitting processed data on another terminal or screen.

[1467] A "system" is a set of devices or programs in which multiple means operate in cooperation with each other.

[1468] The present invention provides a fair customer service experience in a virtual store based on an online interview system with gender-neutral voice conversion and emotion recognition functions. In this embodiment, a user, a terminal, and a server work together.

[1469] Processing on the server

[1470] The server receives the user's voice stream and converts it into a gender-neutral voice. This voice is generated by applying a voice conversion algorithm to ensure a gender-neutral voice. The server then uses an emotion engine to recognize emotions from the user's voice stream. This emotion data plays an important role in the customer service system of the virtual store.

[1471] Processing on the device

[1472] The device acquires a video stream from the user's camera and processes each frame in this video stream in real time. It uses a face detection algorithm such as OpenCV to detect the user's face in each frame. It then superimposes a neutral avatar image on the detected face. The device then analyzes the video stream and uses an emotion engine to recognize emotions from the user's facial expressions. It also has the ability to change the avatar's facial expression based on the recognized emotion.

[1473] User processing

[1474] The user logs into the virtual store customer service system and grants permission to use the camera and microphone. This allows the user's voice and video to be captured. The user's voice is converted to a gender-neutral voice, and a gender-neutral avatar is displayed in the video. Furthermore, the user's emotions are recognized, and this information is conveyed to the store staff in real time. For example, if the user is excited about a product, the customer service system can use this information to make appropriate product introductions and suggestions, thereby increasing the user's desire to purchase.

[1475] Hardware and Software

[1476] Hardware: Camera, microphone, PC

[1477] Software: OpenCV, PyAudio, Emotion Recognizer (our own emotion recognition engine), Gender Neutral Voice (our own voice conversion module)

[1478] Specific examples

[1479] When a customer asks, "Tell me about this product," the system responds in a neutral voice, saying, "This product uses the latest technology and features X features." At the same time, if the customer is excited, the system notifies the store staff, who can then provide more detailed information or special offers.

[1480] Prompt Sentence Examples

[1481] "Give us an example of a process that detects faces in every video frame, overlays them with an androgynous avatar, and performs customer emotion recognition."

[1482] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1483] Step 1:

[1484] The user logs in to the virtual store customer service system and grants permission to use the camera and microphone.

[1485] Input: User's camera image and microphone audio

[1486] Action: Allow access to camera and microphone devices

[1487] Output: You can get the user's audio and video streams.

[1488] Step 2:

[1489] The device acquires the video stream from the camera and starts processing it sequentially in real time.

[1490] Input: Video stream from camera

[1491] Processing: Capture video stream frame by frame

[1492] Output: Individual video frames

[1493] Step 3:

[1494] The device uses OpenCV to detect the user's face in each video frame.

[1495] Input: Individual video frames

[1496] Processing: Apply OpenCV face detection algorithm to identify face area

[1497] Output: Video frames where faces are detected and face coordinate information

[1498] Step 4:

[1499] The device overlays a neutral avatar image onto the detected face.

[1500] Input: Coordinate information of face area

[1501] Processing: Overlay the avatar image onto the detected face area

[1502] Output: Video frame with avatar superimposed

[1503] Step 5:

[1504] The user's voice stream is also simultaneously captured by the terminal and sent to the server.

[1505] Input: Audio stream from microphone

[1506] Processing: Capture the audio stream and send it to the server

[1507] Output: The audio stream sent to the server

[1508] Step 6:

[1509] The server receives the audio stream and converts it to a neutral voice.

[1510] Input: The user's audio stream

[1511] Processing: Applying voice conversion algorithms to generate a neutral voice

[1512] Output: Neutral audio stream

[1513] Step 7:

[1514] The server utilizes an emotion engine to recognize the user's emotion from the audio stream.

[1515] Input: Neutral audio stream

[1516] Processing: Applying emotion engine to analyze emotional state from voice data

[1517] Output: Recognized emotion data

[1518] Step 8:

[1519] The recognized emotion data is transmitted to the terminal.

[1520] Input: Recognized emotion data

[1521] Processing: Emotion data is transferred to the device

[1522] Output: Emotion data sent to the device

[1523] Step 9:

[1524] The device analyzes the video stream and recognizes emotions from the user's facial expressions.

[1525] Input: Video frame

[1526] Processing: Apply video stream analysis and emotion engine to recognize emotions from facial expression data

[1527] Output: Emotion data from recognized facial expressions

[1528] Step 10:

[1529] The device changes the avatar's facial expression based on the recognized emotion.

[1530] Input: Emotion data recognized from voice and facial expressions

[1531] Processing: Dynamically change avatar facial expressions

[1532] Output: Video frames with an avatar superimposed with facial expressions corresponding to the emotion.

[1533] Step 11:

[1534] The processed audio and video streams are output in real time to shop staff in the virtual store.

[1535] Input: Neutral audio stream and video frames with avatars superimposed

[1536] Processing: Transfer of audio and video streams

[1537] Output: Audio and video streams displayed on shop staff's terminals

[1538] 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.

[1539] 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.

[1540] 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.

[1541] 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.

[1542] 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.

[1543] 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.

[1544] 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).

[1545] 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.

[1546] 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."

[1547] 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.

[1548] 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).

[1549] 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.

[1550] 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.

[1551] 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.

[1552] 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.

[1553] 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.

[1554] 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.

[1555] 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.

[1556] 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.

[1557] 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.

[1558] 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.

[1559] The following is further disclosed regarding the above embodiment.

[1560] (Claim 1)

[1561] means for receiving a user's voice stream and converting it to a neutral voice;

[1562] means for capturing a video stream of a user and sequentially processing the video stream to detect the user's face;

[1563] means for superimposing a gender-neutral avatar on the detected face;

[1564] The system includes means for outputting the processed audio and video streams.

[1565] (Claim 2)

[1566] The system of claim 1, further comprising applying an algorithm to convert the audio stream to a neutral voice.

[1567] (Claim 3)

[1568] 10. The system of claim 1, further comprising: applying a face detection algorithm to sequentially process the user's video stream to detect faces.

[1569] (Claim 4)

[1570] 10. The system of claim 1, further comprising means for resizing an avatar image to superimpose a gender-neutral avatar on the detected face.

[1571] "Example 1"

[1572] (Claim 1)

[1573] means for receiving user voice data and converting it into a neutral voice;

[1574] means for acquiring video data of a user and sequentially processing the data to detect the user's face;

[1575] means for superimposing a gender-neutral virtual character on the detected face;

[1576] The system includes means for outputting the processed audio and video data.

[1577] (Claim 2)

[1578] The system of claim 1, further comprising applying an algorithm to convert the voice data into a neutral voice.

[1579] (Claim 3)

[1580] 10. The system of claim 1, further comprising: applying a face detection algorithm to sequentially process the user's video data to detect a face.

[1581] "Application Example 1"

[1582] (Claim 1)

[1583] means for receiving a user's voice stream and converting it to a neutral voice;

[1584] means for capturing a video stream of a user and sequentially processing the video stream to detect the user's face;

[1585] means for superimposing a gender-neutral avatar on the detected face;

[1586] means for outputting the processed audio and video streams to an interactive system of the autonomous vehicle;

[1587] means for outputting a user's neutral voice to an in-vehicle speaker;

[1588] means for displaying a gender-neutral avatar on an in-vehicle display using face detection;

[1589] A system including:

[1590] (Claim 2)

[1591] The system of claim 1, further comprising applying an algorithm to convert the audio stream to a neutral voice.

[1592] (Claim 3)

[1593] 10. The system of claim 1, further comprising: applying a face detection algorithm to sequentially process the user's video stream to detect faces.

[1594] "Example 2: Combining Emotion Engines"

[1595] (Claim 1)

[1596] means for receiving a user's voice stream and converting it to a neutral voice;

[1597] means for analyzing the audio stream and recognizing emotions;

[1598] means for outputting the recognized emotion data;

[1599] means for capturing a video stream of a user and sequentially processing the video stream to detect the user's face;

[1600] means for superimposing a gender-neutral avatar on the detected face;

[1601] means for analyzing the video stream and recognizing emotions from the user's facial expressions;

[1602] means for varying the facial expression of the avatar based on the recognized emotion;

[1603] A means for outputting the processed audio and video streams

[1604] A system including:

[1605] (Claim 2)

[1606] The system of claim 1, further comprising applying an algorithm to convert the audio stream to a neutral voice.

[1607] (Claim 3)

[1608] 10. The system of claim 1, further comprising: applying a face detection algorithm to sequentially process the user's video stream to detect faces.

[1609] "Application example 2 when combining emotion engines"

[1610] (Claim 1)

[1611] means for receiving a user's voice stream and converting it to a neutral voice;

[1612] means for capturing a video stream of a user and sequentially processing the video stream to detect the user's face;

[1613] means for superimposing a gender-neutral avatar on the detected face;

[1614] means for recognizing emotions from a user's audio and video streams;

[1615] means for varying the facial expression of the avatar based on the recognized emotion;

[1616] The system includes means for outputting the processed audio and video streams.

[1617] (Claim 2)

[1618] The system of claim 1, further comprising applying an algorithm to convert the audio stream to a neutral voice.

[1619] (Claim 3)

[1620] 10. The system of claim 1, further comprising: applying a face detection algorithm to sequentially process the user's video stream to detect faces. [Explanation of symbols]

[1621] 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 receiving a user's voice stream and converting it to a neutral voice; means for capturing a video stream of a user and sequentially processing the video stream to detect the user's face; means for superimposing a gender-neutral avatar on the detected face; The system includes means for outputting the processed audio and video streams.

2. The system of claim 1 , further comprising: applying an algorithm to convert the audio stream to a neutral sound.

3. The system of claim 1 , further comprising: applying a face detection algorithm to sequentially process the user's video stream to detect faces.

4. 2. The system of claim 1, further comprising means for resizing an avatar image to superimpose a gender-neutral avatar on the detected face.

Images