System

Abstract

A system is provided.SOLUTION: A system comprising: means for extracting and converting audio data into text; means for analyzing video data to detect slide content and pointer movement; means for identifying a speaker using facial recognition techniques; means for integrating audio data and video data to generate minutes; and means for storing and making searchable the generated minutes.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] Conventional minutes-taking systems mainly use only audio data, making it difficult to grasp the details of the meeting content using materials or identify speakers. Furthermore, the minutes generated in this way are often incomplete, making it difficult to review the content later. The present invention solves these problems by providing a system that comprehensively records the entire meeting and generates accurate and detailed minutes. [Means for solving the problem]

[0005] This invention provides a system that includes means for extracting audio data and converting it into text, means for analyzing video data to detect slide content and pointer movement, means for identifying speakers using face recognition technology, means for generating minutes by integrating audio and video data, and means for saving and making the generated minutes searchable. This allows users to accurately grasp the overall picture of a meeting and the detailed remarks made by each speaker, and easily review and reuse the information later.

[0006] "Audio data" refers to audio information generated during a meeting or conversation that has been recorded in digital format.

[0007] "Video data" refers to visual information such as videos and slides displayed during meetings and conversations, recorded in digital format.

[0008] A "voice recognition module" is a software or hardware technology for converting voice data into text data.

[0009] A "frame" is an individual still image that makes up video data, and is the basic unit of moving images.

[0010] "Facial recognition technology" is a technology for detecting a person's face from video data and identifying its features.

[0011] "Pointer" refers to the cursor or light spot used to indicate a point of interest in a presentation or slide show.

[0012] "Minutes" are a written record of the contents of a meeting, decisions made, and statements made by speakers.

[0013] "Searchable" refers to the ability to efficiently find specific information from stored data.

[0014] "Module" refers to a highly self-contained piece of software or hardware designed to perform a specific function.

[0015] "Text data" refers to data expressed in characters that has been converted by a voice recognition module. [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] This invention is a system that analyzes the audio and video of a meeting and automatically generates detailed minutes. This system processes audio and video data, accurately capturing the content of materials and participants' remarks, and creates minutes.

[0038] System Program Overview

[0039] The system begins by having a user upload a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers. This data is then integrated to generate accurate minutes. The generated minutes are then stored on the server and made available in a searchable format.

[0040] Audio data processing

[0041] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[0042] Video data processing

[0043] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[0044] Identifying the speaker

[0045] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[0046] Generate meeting minutes

[0047] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in the specified format, which contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[0048] Providing generated minutes

[0049] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search for minutes based on "what a specific speaker said" or "a specific slide from the meeting."

[0050] Specific examples

[0051] For example, a user uploads a video of a "project progress meeting" to the server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify speakers, and this information is integrated to generate minutes. The generated minutes include details such as "Yamada spoke about important point A while explaining slide 3." These minutes are stored on the server and can be searched and viewed by users later.

[0052] This system allows detailed recording of meeting contents, allowing users to easily check and reuse the information later.

[0053] The processing flow will be explained below.

[0054] Step 1:

[0055] The user uploads the video file of the conference from the terminal to the server.

[0056] Step 2:

[0057] The server receives the uploaded video file and stores it in a temporary storage area.

[0058] Step 3:

[0059] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[0060] Step 4:

[0061] The server uses a voice recognition module (for example, Google® Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[0062] Step 5:

[0063] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[0064] Step 6:

[0065] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[0066] Step 7:

[0067] The server uses a face recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected face data is matched with the audio data.

[0068] Step 8:

[0069] The server combines the results of speech recognition, video analysis, and speaker identification to generate a draft of the minutes, which includes the text data of each statement, slide content, pointer movement, and speaker information.

[0070] Step 9:

[0071] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[0072] Step 10:

[0073] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[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] Accurately analyzing the audio and video of a meeting and automatically generating detailed minutes is a time-consuming and labor-intensive process. Furthermore, it is often difficult to accurately identify speakers and record slide content. Furthermore, it is difficult to quickly search through the generated minutes to find specific information.

[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 extracting audio data and converting it to text, means for dividing and analyzing video data by frame to detect slide content and pointer movement, means for identifying speakers using face recognition technology, means for generating minutes by integrating the results of the audio recognition, video analysis, and speaker identification, and means for saving the generated minutes and providing them in a searchable format, thereby enabling accurate recording of the meeting contents and rapid subsequent retrieval.

[0079] "Audio data" refers to digital data that is a recording of what is said during a meeting or the like.

[0080] "Video data" refers to digital data in the form of moving images captured during a conference.

[0081] A "voice recognition module" is a software or hardware component for converting voice data into text data.

[0082] A "frame" is a unit that refers to each still image that makes up video data.

[0083] A "video analysis module" is a software or hardware component that analyzes video data to detect slide content and pointer movement.

[0084] "Slide content" refers to the information on presentation slides displayed during a conference.

[0085] "Pointer movement" refers to the movement of the pointer or mouse on the slides during the meeting.

[0086] "Facial recognition technology" is a technology for identifying and recognizing individual faces from video data.

[0087] "Speaker" refers to a person who speaks during a meeting.

[0088] A "minutes" is a document that records in detail the progress of a meeting, what was said, the contents of slides, etc.

[0089] A "searchable format" is a data format that is indexed so that a user can later search for specific content.

[0090] The present invention is a system for automatically generating detailed meeting minutes by analyzing audio and video of a meeting. The system uses multiple hardware and software modules to process audio and video data and generate accurate meeting minutes.

[0091] Basic configuration

[0092] Audio data processing

[0093] The system begins operation when a user uploads a video file of a meeting to the server. The server first extracts audio data from the video file using a tool such as FFmpeg. This extracted audio data is then sent to a speech recognition module (e.g., Google Cloud Speech-to-Text API) and converted into text data. This text data is then used to record what was said by the meeting participants.

[0094] Video data processing

[0095] Next, the server splits the video file into frames using a tool such as FFmpeg. Each frame is then analyzed using a video analysis module such as OpenCV. This analysis detects the slide content and pointer movement.

[0096] Identifying the speaker

[0097] Facial recognition technology (e.g., Amazon Rekognition) is used to analyze the video data. The server uses the video data to detect the faces of each participant and identifies the speaker by matching them with the audio data. This makes it clear which participant made each comment.

[0098] Generate meeting minutes

[0099] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in a specified format (e.g., Markdown or PDF). These minutes contain detailed records of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[0100] Providing generated minutes

[0101] The generated minutes are saved on the server. The saved minutes are indexed using ElasticSearch (registered trademark) and provided in a format that makes it easy for users to search later. Users can access the minutes using a web browser or a dedicated app, and search and view them based on "the content of a specific speaker's remarks" or "specific slides."

[0102] Specific examples

[0103] For example, a user uploads a video file of a "project progress meeting" to the server. This video file can be in supported formats such as MP4 or AVI. The server first extracts the audio data and converts it to text using the Google Cloud Speech-to-Text API. Next, the server divides the video data into frames and analyzes the slide content and pointer movement using OpenCV. It identifies speakers using facial recognition technology (e.g., Amazon Rekognition), and combines this data to generate minutes. The generated minutes include details such as "Participant A spoke about an important point while explaining slide 3." The minutes are stored on the server and can be searched and viewed later by users.

[0104] Prompt Sentence Examples

[0105] I have uploaded a video of a "Project Status Meeting." Please generate minutes of this meeting. The minutes should include the content of each slide and what each speaker said.

[0106] This system automatically and accurately records detailed meeting content, allowing users to quickly search for and check the information they need later.

[0107] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0108] Step 1:

[0109] The user uploads the video file of the conference to the server. At this stage, the video file is sent to the server and saved in the specified directory. The input is the video file provided from the user's device, and the output is the video file saved on the server side.

[0110] Step 2:

[0111] The server receives the video file and extracts the audio data. Using a tool such as FFmpeg, it separates the audio track from the video file. The input is the video file saved in step 1, and the output is an audio file temporarily stored on the server.

[0112] Step 3:

[0113] The server sends the audio data to the speech recognition module. The audio file is converted into text data using a method such as the Google Cloud Speech-to-Text API. During this process, the speech recognition module is called by the server. The input is the audio file, and the output is text data.

[0114] Step 4:

[0115] The server receives the text data from the speech recognition module and stores it as the speech content. This allows each speech made during the meeting to be recorded in text format. The input is the text data of the speech recognition results, and the output is the storage of the text data as the speech content.

[0116] Step 5:

[0117] The server splits the video file into frames. FFmpeg is used to convert the video into frames, which are a series of still images. The input is the video file saved in step 1, and the output is an image file split into frames.

[0118] Step 6:

[0119] The server sends the image files divided into frames to a video analysis module for analysis. Using a video analysis module such as OpenCV, slide content and pointer movement are detected from each frame. The input is the frame image file, and the output is the analysis results: slide content and pointer movement information.

[0120] Step 7:

[0121] The server uses facial recognition technology to identify the speaker. Using Amazon Rekognition or similar, it detects the face of each participant from the video data and compares it with the audio data. The input is the frame image file and audio data, and the output is data identifying the speaker.

[0122] Step 8:

[0123] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes. The minutes are created in the specified format (Markdown, PDF, etc.). The input is the data obtained at each processing step (text data, analysis results, speaker information), and the output is the completed minutes file.

[0124] Step 9:

[0125] The generated minutes are stored on a server and provided in a searchable format. They are indexed using a search engine such as Elasticsearch, allowing users to easily access and search them later. The input is the generated minutes file, and the output is an indexed database.

[0126] Step 10:

[0127] Users search and view meeting minutes using a web browser or a dedicated app. The server returns indexed meeting minutes in response to the user's request. The input is the user's search query, and the output is the relevant part of the meeting minutes as a search result.

[0128] (Application example 1)

[0129] 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."

[0130] Conventional meeting minutes systems require manual processing of audio and video data to create minutes, requiring a great deal of time and effort. Similar problems existed in online lectures and seminars, making it difficult to accurately record the content of lectures and details of Q&A sessions. Furthermore, searching this data later required a great deal of effort, preventing sufficient reuse of information.

[0131] 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.

[0132] In this invention, the server includes means for extracting audio data and converting it to text, means for analyzing video data to detect slide content and pointer movement, means for identifying speakers using facial recognition technology, means for integrating audio and video data to generate minutes, means for saving the generated minutes and making them searchable, means for extracting audio data from video, means for transmitting the extracted audio data to a voice recognition system and converting it to text data, and means for recording the content of lectures and question times. This allows the content of lectures and meetings to be automatically and accurately recorded, making it possible to easily search and reuse the information later.

[0133] "Audio data" refers to audio data recorded during a conference or lecture.

[0134] "Text data" is character string data converted from voice data using voice recognition technology.

[0135] "Video data" refers to video frame data recorded during a conference or lecture.

[0136] "Slide content" refers to information on presentation materials used during lectures and meetings.

[0137] "Pointer movement" refers to the movement trajectory of the cursor or pointer indicated on the screen during a presentation.

[0138] "Facial recognition technology" is a technology that identifies the face of a specific person from video data.

[0139] "Speaker" refers to the person who made a statement at a meeting or lecture.

[0140] "Integration" means combining individual audio data, video data, and speaker information into a single minutes.

[0141] "Minutes" are documents that record the progress of a meeting or lecture, what was said, slide information, etc.

[0142] "Searchable" means that stored information can be easily found based on specified criteria.

[0143] A "video" is a data file in which video and audio are recorded together.

[0144] A "voice recognition system" is a system that analyzes voice data and converts it into text data.

[0145] "Lecture content" refers to a series of statements, slides, and content intended for educational or explanatory purposes.

[0146] "Question windows" refer to specific time slots during lectures or meetings when questions arise.

[0147] This invention is a system that analyzes audio and video data from lectures and seminars and automatically generates detailed minutes. This system is realized using hardware such as a server, smartphone, and head-mounted display, and software such as a voice recognition module, video analysis module, and facial recognition technology.

[0148] System Overview

[0149] The system starts by a user uploading a video of a lecture or meeting to the server, which receives the video and performs the following steps:

[0150] 1. Extract audio data from video.

[0151] 2. The extracted voice data is sent to a voice recognition system and converted into text data.

[0152] 3. Analyze the video data to detect the slide content and pointer movement.

[0153] 4. Use facial recognition technology to identify speakers.

[0154] 5. Integrate these data to generate detailed meeting minutes.

[0155] 6. The generated minutes are saved on the server and made searchable.

[0156] Hardware and software used

[0157] Hardware:

[0158] Smartphones and head-mounted displays: Used to record and upload videos.

[0159] Server: Used to analyze data and generate transcripts.

[0160] software:

[0161] Python: Implementation of the main program and use of libraries.

[0162] cv2 (OpenCV): Video data analysis and frame segmentation.

[0163] speech_recognition: Converting speech data to text.

[0164] face_recognition: Application of facial recognition technology.

[0165] moviepy: Extracting audio data from videos.

[0166] transformers (Hugging Face): Speech recognition using the Wav2Vec2 model.

[0167] Processing flow

[0168] 1. Extracting audio data:

[0169] The server extracts the audio data from the uploaded video, which it does using the moviepy library.

[0170] 2. Speech Recognition:

[0171] The extracted audio data is converted into text data by a speech recognition system using the speech_recognition library. At this stage, the lecture content and speech content are obtained as text information.

[0172] 3. Video Data Analysis:

[0173] The server splits the video data into frames, detects the slide content and pointer movement using the OpenCV library, and identifies the speaker by recognizing the faces in each frame using the face_recognition library.

[0174] 4. Integration and Transcription:

[0175] The speech recognition results, video analysis results, and speaker information obtained at each step are integrated to generate minutes in the specified format. For example, the minutes will include information such as "Speaker A spoke about important point A while explaining slide 3."

[0176] 5. Archive and search meeting minutes:

[0177] The generated minutes are saved on the server for users to view and search later. Users can easily search for minutes based on "the content of a specific speaker" or "a specific slide."

[0178] Specific examples

[0179] For example, when a user uploads a video of an online lecture to the server, the system automatically analyzes the audio and video and generates detailed minutes that record the lecture content and the time periods when questions were asked. These minutes include detailed information such as, "Speaker A began explaining slide 1 immediately after the lecture began."

[0180] Prompt Sentence Examples

[0181] "Please convert the following audio data into text. Link to audio data: {URL of audio data}. Please describe the audio in chronological order and include speaker information if available."

[0182] As described above, the present invention makes it possible to automatically record the contents of lectures and meetings, and to easily search and reuse them later.

[0183] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0184] Step 1:

[0185] A user uploads video files of lectures or meetings to the server. The user selects the video files using their device and sends them to a specific upload destination on the server. The server stores the received video files and prepares them for further processing.

[0186] Input: Video file uploaded by user.

[0187] Output: Video file saved on the server.

[0188] Step 2:

[0189] The server extracts the audio data from the stored video file. For this process, it uses the moviepy library. The server reads the video file, separates the audio track, and saves it as a new audio file.

[0190] Input: Video file.

[0191] Output: Audio file.

[0192] Step 3:

[0193] The server sends the extracted audio file to a speech recognition module and converts it into text data. This process uses the speech_recognition library. The server reads the audio file and passes it through a speech recognition model to obtain text information.

[0194] Input: Audio file.

[0195] Output: Text data.

[0196] Step 4:

[0197] The server splits the video data from the video file into frames and analyzes them. This process uses the OpenCV library. The server splits the video file into frames and identifies the slide content, pointer movement, and speaker for each frame.

[0198] Input: Video file.

[0199] Output: Video data for each frame, analysis results (slide content, pointer movement, speaker information).

[0200] Step 5:

[0201] The server identifies the speaker based on the facial information acquired for each frame. Using the face_recognition library, the server compares the facial data in each frame with known facial data to identify the speaker.

[0202] Input: Face data for each frame.

[0203] Output: Identified speaker information.

[0204] Step 6:

[0205] The server then combines the acquired text data, slide content, pointer movement, and speaker information to generate detailed minutes. By combining each piece of data, the minutes are compiled in a format that users can easily review later.

[0206] Input: text data, slide content, pointer movement, speaker information.

[0207] Output: Detailed meeting transcript.

[0208] Step 7:

[0209] The server stores the generated minutes and provides them to users in a searchable format. The minutes are stored in a database and indexed so that users can easily search for specific criteria.

[0210] Input: Detailed minutes.

[0211] Output: Saved, searchable meeting minutes.

[0212] By using the above processing steps, the user can automatically record the contents of a lecture or conference, and later easily check the overall progress or the content of specific comments.

[0213] 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.

[0214] This invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates minutes of the meeting with an emotion engine that recognizes the emotions of users. This system processes audio and video data and utilizes the emotion engine to record the emotions of participants and reflect them in the minutes, making it possible to comprehensively understand the detailed situation of the meeting.

[0215] System Program Overview

[0216] The system begins by a user uploading a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers and an emotion engine to recognize participants' emotions. This data is then integrated to generate accurate minutes, which are then stored on the server and made available in a searchable format.

[0217] Audio data processing

[0218] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[0219] Video data processing

[0220] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[0221] Identifying the speaker

[0222] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[0223] Emotion recognition

[0224] Using the emotion engine, a distinctive feature of the present invention, the server recognizes participants' emotions from the video and audio data. The emotion engine uses techniques such as facial expression analysis and audio tone analysis to identify the emotional state (e.g., joy, anger, sadness, etc.) of each speaker. The recognized emotion information is saved as text data.

[0225] Generate meeting minutes

[0226] The server combines the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes in a specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[0227] Providing generated minutes

[0228] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[0229] Specific examples

[0230] For example, a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify the speaker, and an emotion engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Yamada spoke about important point A while explaining slide 3, and at the same time, Yamada seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[0231] This system not only records the details of the meeting, but also grasps the emotional state of the participants, allowing for deeper understanding and analysis.

[0232] The processing flow will be explained below.

[0233] Step 1:

[0234] The user uploads the video file of the conference from the terminal to the server.

[0235] Step 2:

[0236] The server receives the uploaded video file and stores it in a temporary storage area.

[0237] Step 3:

[0238] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[0239] Step 4:

[0240] The server uses a speech recognition module (e.g., Google Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[0241] Step 5:

[0242] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[0243] Step 6:

[0244] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[0245] Step 7:

[0246] The server uses a facial recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected facial data is matched with the audio data.

[0247] Step 8:

[0248] The server uses an emotion engine to recognize participants' emotions from video and audio data. The emotion engine analyzes facial expressions and audio tones, and stores the emotional state as text data.

[0249] Step 9:

[0250] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate a draft of the minutes, which includes text data for each statement, slide content, pointer movement, speaker information, and emotion information.

[0251] Step 10:

[0252] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[0253] Step 11:

[0254] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[0255] Example 2

[0256] 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."

[0257] Conventional meeting recording systems require a great deal of time and effort to process audio and video data separately and generate minutes manually. Furthermore, there is no way to grasp the emotional state of participants, making it difficult to understand the detailed situation of the meeting or the emotions of the participants. Furthermore, the generated minutes are not easily searchable or categorized, making them difficult to reference later.

[0258] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for extracting audio data and converting it into text, means for analyzing video data to detect visual information and indicator movements, means for identifying the caller using biometric recognition technology, means for recognizing an emotional state from the audio data and video data using a recognition engine, means for generating minutes by integrating the audio data, video data, caller information, and emotional information, and means for saving the generated minutes and making them categorizable. This enables audio and video to be processed in an integrated manner, automatically generating minutes that also include the caller's emotional state, and enabling efficient search and reference.

[0259] "Audio data" refers to recordings of sounds collected during meetings, conversations, etc.

[0260] "Text" is character information generated by processing audio data.

[0261] "Video data" refers to video information containing visual content such as meetings and presentations.

[0262] "Visual information" refers to information such as slide content and screen display analyzed from video data.

[0263] "Pointer movement" refers to interactive elements such as pointer or hand movements detected within the video data.

[0264] "Biometric recognition technology" is a technology that identifies individuals based on video data, such as facial recognition.

[0265] A "caller" is a person participating in a meeting or conversation.

[0266] A "cognitive engine" is a technology that analyzes audio and video data to recognize emotional states and other perceptual information.

[0267] "Emotional state" is information that indicates the psychological reactions and emotions (joy, anger, sadness, etc.) of the speaker or participants.

[0268] "Integration" is the process of combining various data to create a single, continuous record.

[0269] A "minutes" is a document that summarizes the contents of a meeting, what was said, the identity of the speakers, and their emotional state.

[0270] "Classifiable" means that the generated minutes are organized according to specific criteria or categories so that they can be easily searched and referenced.

[0271] A "system" is a structure having a series of integrated functions that includes all of the above means.

[0272] The present invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates meeting minutes with a recognition engine that recognizes user emotions. An embodiment of the present invention will be described in detail below.

[0273] The system starts by having a user upload a video of a meeting to the server, which receives the video and processes it in the following steps:

[0274] First, the server uses a library such as "FFmpeg" to extract audio data from the video. The extracted audio data is sent to a "speech recognition module" (general name) and converted into text data. For this purpose, a "speech recognition API" (Google Cloud Speech-to-Text API) or similar is used. This text data is then used to record what the participants are saying.

[0275] Next, the server divides the video into frames and processes the video data. A "video analysis module" (OpenCV) is used to divide the frames. The video analysis module analyzes each frame and detects the contents of the slides and the movement of the pointer. This information is used to record the details of the materials used in the meeting. In addition, "biometric recognition technology" (facial recognition technology) is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker.

[0276] An important feature is that the server uses a recognition engine, which includes technologies such as facial expression analysis and voice tone analysis (IBM Watson (registered trademark) Tone Analyzer). Using these technologies, the server recognizes the emotional state (e.g., joy, anger, sadness, etc.) of each speaker from the video and voice data. The recognized emotional information is saved as text data.

[0277] Once this data is collected, the server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition, and generates minutes in the specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[0278] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[0279] As a concrete example, imagine a scenario where a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text using a "speech recognition API." Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Biometric recognition technology is used to identify the speaker, and a recognition engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Person A spoke about important point A while explaining slide 3, and at the same time, Person A seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[0280] An example prompt might be, "Upload a video of a project status meeting and automatically generate minutes. Please also record what Person A said and their emotions at the time."

[0281] The above is a specific description of the embodiment of the present invention. The present invention not only records the details of the meeting but also grasps the emotional states of the participants, making it possible to comprehensively understand and analyze the detailed situation of the meeting.

[0282] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0283] Step 1:

[0284] A user uploads a video of the meeting.

[0285] The user selects the video file of the "Project Progress Meeting" from their device and sends it to the server through the system's upload interface. As input, the video file of the meeting is passed from the device to the server. As output, the video file is saved on the server.

[0286] Step 2:

[0287] The server extracts the voice data and sends it to a voice recognition module.

[0288] The server analyzes the uploaded video file and extracts the audio data using a library such as FFmpeg. This audio data is then sent to a speech recognition API. The audio data extracted from the video file is used as input. The audio data is converted to text as output.

[0289] Step 3:

[0290] The server divides the video data into frames and processes them in a video analysis module.

[0291] The server splits the video data into frames using OpenCV. For each frame, it uses OCR technology to recognize the slide content and analyzes the pointer, hand, and other indicator movements. The video data is used as input. Data about the slide content and indicator movements is generated as output.

[0292] Step 4:

[0293] The server uses facial recognition technology to identify the speaker.

[0294] The server detects facial images in each frame of video data and identifies the speaker using facial recognition technology such as Amazon Rekognition. The detected facial images are then matched with the audio data. The facial images and audio data are used as input. The output is data that identifies the speaker.

[0295] Step 5:

[0296] The server uses an emotion engine to recognize the emotions of the participants.

[0297] The server uses an emotion engine to analyze the video and audio data and recognize the speaker's emotional state. This process uses facial expression analysis and tone analysis techniques. The video and audio data are used as input. The output is an emotion recognition result (e.g., joy, anger, sadness).

[0298] Step 6:

[0299] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes.

[0300] The server integrates the data collected so far and generates minutes in the specified format. These minutes include detailed records of the progress of the meeting, the contents of the materials, and the remarks and emotions of the speakers. Various analysis and recognition results are used as input. The integrated minutes data is generated as output.

[0301] Step 7:

[0302] The server provides the generated minutes.

[0303] The generated minutes are stored on a server and provided in a format that users can access later for searching and viewing. Specifically, users can use the search interface to search for "emotional reactions to a specific slide in a meeting," "the content of a specific speaker's remarks and their emotions at the time," and so on. The user's search query is used as input, and the relevant minutes data is provided as output.

[0304] The above is the specific processing flow of the system program divided into specific processing steps, and the details of each process accompanied by input and output.

[0305] (Application example 2)

[0306] 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."

[0307] Conventional meeting recording systems convert audio data into text and analyze video data, but lack the functionality to monitor participants' emotions and work environment in real time. This makes it difficult to understand the details of the meeting content or to properly grasp the fatigue and stress levels of workers. Furthermore, there is a need for a system that can respond in real time to changes in the work environment.

[0308] 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 extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movement, means for identifying the speaker using face recognition technology, means for recognizing the emotions of participants using emotion recognition technology, means for integrating audio data and video data to generate minutes, means for saving the generated minutes and making them searchable, and means for monitoring the work environment in real time and detecting the fatigue and stress of workers. This makes it possible to not only record detailed information about meetings and work content, but also to monitor the emotions of participants and the work environment in real time.

[0309] "Voice data" refers to voice information such as statements and instructions that occur during meetings or work.

[0310] "Means for converting to text" refers to the technology or equipment that converts voice data into text information.

[0311] "Video data" refers to video information captured by cameras or recording devices.

[0312] "Slide content" refers to the information on slides used in presentations or as meeting materials.

[0313] "Indicator movement" refers to the movement of a pointer, mouse cursor, or other device on the screen to indicate instructions or a point of interest.

[0314] "Facial recognition technology" refers to technology for recognizing and identifying specific individuals from video data.

[0315] "Means for identifying the speaker" refers to technologies and methods for determining who is speaking based on audio and video data.

[0316] "Emotion recognition technology" refers to technology for identifying an individual's emotional state (joy, anger, sadness, etc.) from audio and video data.

[0317] "Means for recognizing emotions of participants" refers to a technique or method for determining the emotions of participants in a meeting or workplace using emotion recognition technology.

[0318] "Means for generating minutes" refers to technologies and methods that integrate audio data, video data, emotional information, etc. to automatically generate documents that record the content of meetings and work.

[0319] "Means for storing and making searchable minutes" refers to the technology and methods for electronically storing the generated minutes in a database so that they can be searched and viewed at a later date.

[0320] "Means for monitoring the work environment in real time" refers to technologies and methods for instantly monitoring the progress and environment of work in factories and workplaces.

[0321] "Means for detecting worker fatigue and stress" refers to technologies and methods for assessing the health and mental state of workers using emotion recognition technology or other sensor technology.

[0322] The present invention relates to a work monitoring and quality control system using factory robots. The system includes means for extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movements, means for identifying speakers using facial recognition technology, means for recognizing participants' emotions using emotion recognition technology, means for integrating audio data and video data to generate minutes of meetings, means for storing the generated minutes and making them searchable, and means for monitoring the work environment in real time to detect worker fatigue and stress.

[0323] System Program Overview

[0324] The server collects video and audio data from within the factory using high-resolution cameras and microphones. The collected audio data is converted into text using a speech recognition module (e.g., IBM Watson Speech to Text). The video data is divided into frames and analyzed using OpenCV. The speaker is identified using slide content, indicator movements, and facial recognition technology.

[0325] Using emotion recognition technology (e.g., Affectiva SDK), the emotions of workers are recognized from video and audio data. This allows participants' emotional information to be saved as text data. The results of audio recognition, video analysis, speaker identification, and emotion recognition are integrated to generate minutes in the specified format.

[0326] Hardware and Software Use

[0327] The system uses a high-resolution camera, microphone, high-performance CPU and GPU, and utilizes software such as TENSORFLOW (registered trademark) (machine learning library), OpenCV (image processing library), IBM Watson Speech to Text (voice recognition), and Affectiva SDK (emotion recognition).

[0328] Specific examples

[0329] For example, if a factory worker says, "I can't assemble this part properly," the voice recording is collected and converted into text. Emotion recognition technology recognizes that the worker is feeling "frustrated." This information is collected and analyzed as follows:

[0330] Audio data:

[0331] "This part doesn't assemble properly" -> recognized_text: "This part doesn't assemble properly"

[0332] Emotional Data:

[0333] recognized_emotion: "frustration"

[0334] We also provide an example of input to a generative AI model using the following prompt sentence:

[0335] "Generate a monitoring report for the factory. Analyze the voice and emotions of the workers and incorporate them into the report."

[0336] This allows the system to record detailed factory working conditions as well as monitor workers' emotions and health in real time, allowing appropriate responses to be taken.

[0337] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0338] Step 1:

[0339] The server uses microphones to collect voice data from within the factory. This allows it to obtain voice information such as conversations and instructions within the factory. Specifically, the analog voice signals collected by the microphones are converted into digital data and input into a voice recognition module. The input is voice data, and the output is digitized voice data.

[0340] Step 2:

[0341] The server uses a speech recognition module (e.g., IBM Watson Speech to Text) to convert the collected voice data into text. The input is digital voice data, and the output is text data. Specifically, the voice signal is converted into acoustic features, which are then input into a model to generate a corresponding string of characters.

[0342] Step 3:

[0343] The server uses a high-resolution camera to collect video data within the factory, visually recording the progress and environment of work. The input is visual information and the output is video data. The video data collected by the camera is divided into frames and input into the image processing module.

[0344] Step 4:

[0345] The server divides the video data into frames and analyzes them using OpenCV. The input is the video data, and the output is analyzed data on the slide content and indicator movement. Specifically, image processing is performed on each frame to extract the characteristics of the slides and indicators.

[0346] Step 5:

[0347] The server uses facial recognition technology (e.g., OpenCV's facial recognition module) to identify speakers from video data. The input is video data, and the output is data on the identified speaker. Specifically, it detects facial features in the video frame and compares them with an existing database to identify individuals.

[0348] Step 6:

[0349] The server uses emotion recognition technology (e.g., Affectiva SDK) to recognize participants' emotions from the collected audio and video data. The input is audio and video data, and the output is emotional state data. Specifically, it performs voice tone analysis and facial expression analysis to identify emotional levels.

[0350] Step 7:

[0351] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to automatically generate minutes in a specified format. The input is a dataset of each analysis result, and the output is the integrated minutes. Specifically, each piece of data is organized chronologically and a document is generated in a specified format.

[0352] Step 8:

[0353] The server stores the generated minutes in a database and provides them in a format that can be searched later. The input is the generated minutes, and the output is a searchable database. Specifically, the minutes data is indexed and stored in the database, making it searchable later using queries.

[0354] Step 9:

[0355] The server monitors the work environment in real time and detects the fatigue and stress levels of workers. The input is video, audio, and emotional data, and the output is data on the fatigue and stress levels of workers. Specifically, it analyzes continuously collected data and generates an alert if a certain threshold is exceeded.

[0356] In this way, it is possible to monitor participants' emotions and work environment in real time, along with detailed records of meetings and work content, and take appropriate measures.

[0357] 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.

[0358] 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.

[0359] 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.

[0360] [Second embodiment]

[0361] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0362] 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.

[0363] 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).

[0364] 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.

[0365] 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.

[0366] 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).

[0367] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0368] 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.

[0369] 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.

[0370] 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.

[0371] 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.

[0372] 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."

[0373] This invention is a system that analyzes the audio and video of a meeting and automatically generates detailed minutes. This system processes audio and video data, accurately capturing the content of materials and participants' remarks, and creates minutes.

[0374] System Program Overview

[0375] The system begins by having a user upload a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers. This data is then integrated to generate accurate minutes. The generated minutes are then stored on the server and made available in a searchable format.

[0376] Audio data processing

[0377] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[0378] Video data processing

[0379] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[0380] Identifying the speaker

[0381] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[0382] Generate meeting minutes

[0383] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in the specified format, which contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[0384] Providing generated minutes

[0385] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search for minutes based on "what a specific speaker said" or "a specific slide from the meeting."

[0386] Specific examples

[0387] For example, a user uploads a video of a "project progress meeting" to the server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify speakers, and this information is integrated to generate minutes. The generated minutes include details such as "Yamada spoke about important point A while explaining slide 3." These minutes are stored on the server and can be searched and viewed by users later.

[0388] This system allows detailed recording of meeting contents, allowing users to easily check and reuse the information later.

[0389] The processing flow will be explained below.

[0390] Step 1:

[0391] The user uploads the video file of the conference from the terminal to the server.

[0392] Step 2:

[0393] The server receives the uploaded video file and stores it in a temporary storage area.

[0394] Step 3:

[0395] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[0396] Step 4:

[0397] The server uses a speech recognition module (e.g., Google Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[0398] Step 5:

[0399] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[0400] Step 6:

[0401] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[0402] Step 7:

[0403] The server uses a face recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected face data is matched with the audio data.

[0404] Step 8:

[0405] The server combines the results of speech recognition, video analysis, and speaker identification to generate a draft of the minutes, which includes the text data of each statement, slide content, pointer movement, and speaker information.

[0406] Step 9:

[0407] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[0408] Step 10:

[0409] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[0410] Example 1

[0411] 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."

[0412] Accurately analyzing the audio and video of a meeting and automatically generating detailed minutes is a time-consuming and labor-intensive process. Furthermore, it is often difficult to accurately identify speakers and record slide content. Furthermore, it is difficult to quickly search through the generated minutes to find specific information.

[0413] 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.

[0414] In this invention, the server includes means for extracting audio data and converting it to text, means for dividing and analyzing video data by frame to detect slide content and pointer movement, means for identifying speakers using face recognition technology, means for generating minutes by integrating the results of the audio recognition, video analysis, and speaker identification, and means for saving the generated minutes and providing them in a searchable format, thereby enabling accurate recording of the meeting contents and rapid subsequent retrieval.

[0415] "Audio data" refers to digital data that is a recording of what is said during a meeting or the like.

[0416] "Video data" refers to digital data in the form of moving images captured during a conference.

[0417] A "voice recognition module" is a software or hardware component for converting voice data into text data.

[0418] A "frame" is a unit that refers to each still image that makes up video data.

[0419] A "video analysis module" is a software or hardware component that analyzes video data to detect slide content and pointer movement.

[0420] "Slide content" refers to the information on presentation slides displayed during a conference.

[0421] "Pointer movement" refers to the movement of the pointer or mouse on the slides during the meeting.

[0422] "Facial recognition technology" is a technology for identifying and recognizing individual faces from video data.

[0423] "Speaker" refers to a person who speaks during a meeting.

[0424] A "minutes" is a document that records in detail the progress of a meeting, what was said, the contents of slides, etc.

[0425] A "searchable format" is a data format that is indexed so that a user can later search for specific content.

[0426] The present invention is a system for automatically generating detailed meeting minutes by analyzing audio and video of a meeting. The system uses multiple hardware and software modules to process audio and video data and generate accurate meeting minutes.

[0427] Basic configuration

[0428] Audio data processing

[0429] The system begins operation when a user uploads a video file of a meeting to the server. The server first extracts audio data from the video file using a tool such as FFmpeg. This extracted audio data is then sent to a speech recognition module (e.g., Google Cloud Speech-to-Text API) and converted into text data. This text data is then used to record what was said by the meeting participants.

[0430] Video data processing

[0431] Next, the server splits the video file into frames using a tool such as FFmpeg. Each frame is then analyzed using a video analysis module such as OpenCV. This analysis detects the slide content and pointer movement.

[0432] Identifying the speaker

[0433] Facial recognition technology (e.g., Amazon Rekognition) is used to analyze the video data. The server uses the video data to detect the faces of each participant and identifies the speaker by matching them with the audio data. This makes it clear which participant made each comment.

[0434] Generate meeting minutes

[0435] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in a specified format (e.g., Markdown or PDF). These minutes contain detailed records of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[0436] Providing generated minutes

[0437] The generated minutes are saved on the server. The saved minutes are indexed using Elasticsearch and provided in a format that makes them easy for users to search later. Users can access the minutes using a web browser or a dedicated app, and search and view them based on "the content of a specific speaker" or "a specific slide."

[0438] Specific examples

[0439] For example, a user uploads a video file of a "project progress meeting" to the server. This video file can be in supported formats such as MP4 or AVI. The server first extracts the audio data and converts it to text using the Google Cloud Speech-to-Text API. Next, the server divides the video data into frames and analyzes the slide content and pointer movement using OpenCV. It identifies speakers using facial recognition technology (e.g., Amazon Rekognition), and combines this data to generate minutes. The generated minutes include details such as "Participant A spoke about an important point while explaining slide 3." The minutes are stored on the server and can be searched and viewed later by users.

[0440] Prompt Sentence Examples

[0441] I have uploaded a video of a "Project Status Meeting." Please generate minutes of this meeting. The minutes should include the content of each slide and what each speaker said.

[0442] This system automatically and accurately records detailed meeting content, allowing users to quickly search for and check the information they need later.

[0443] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0444] Step 1:

[0445] The user uploads the video file of the conference to the server. At this stage, the video file is sent to the server and saved in the specified directory. The input is the video file provided from the user's device, and the output is the video file saved on the server side.

[0446] Step 2:

[0447] The server receives the video file and extracts the audio data. Using a tool such as FFmpeg, it separates the audio track from the video file. The input is the video file saved in step 1, and the output is an audio file temporarily stored on the server.

[0448] Step 3:

[0449] The server sends the audio data to the speech recognition module. The audio file is converted into text data using a method such as the Google Cloud Speech-to-Text API. During this process, the speech recognition module is called by the server. The input is the audio file, and the output is text data.

[0450] Step 4:

[0451] The server receives the text data from the speech recognition module and stores it as the speech content. This allows each speech made during the meeting to be recorded in text format. The input is the text data of the speech recognition results, and the output is the storage of the text data as the speech content.

[0452] Step 5:

[0453] The server splits the video file into frames. FFmpeg is used to convert the video into frames, which are a series of still images. The input is the video file saved in step 1, and the output is an image file split into frames.

[0454] Step 6:

[0455] The server sends the image files divided into frames to a video analysis module for analysis. Using a video analysis module such as OpenCV, slide content and pointer movement are detected from each frame. The input is the frame image file, and the output is the analysis results: slide content and pointer movement information.

[0456] Step 7:

[0457] The server uses facial recognition technology to identify the speaker. Using Amazon Rekognition or similar, it detects the face of each participant from the video data and compares it with the audio data. The input is the frame image file and audio data, and the output is data identifying the speaker.

[0458] Step 8:

[0459] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes. The minutes are created in the specified format (Markdown, PDF, etc.). The input is the data obtained at each processing step (text data, analysis results, speaker information), and the output is the completed minutes file.

[0460] Step 9:

[0461] The generated minutes are stored on a server and provided in a searchable format. They are indexed using a search engine such as Elasticsearch, allowing users to easily access and search them later. The input is the generated minutes file, and the output is an indexed database.

[0462] Step 10:

[0463] Users search and view meeting minutes using a web browser or a dedicated app. The server returns indexed meeting minutes in response to the user's request. The input is the user's search query, and the output is the relevant part of the meeting minutes as a search result.

[0464] (Application example 1)

[0465] 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."

[0466] Conventional meeting minutes systems require manual processing of audio and video data to create minutes, requiring a great deal of time and effort. Similar problems existed in online lectures and seminars, making it difficult to accurately record the content of lectures and details of Q&A sessions. Furthermore, searching this data later required a great deal of effort, preventing sufficient reuse of information.

[0467] 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.

[0468] In this invention, the server includes means for extracting audio data and converting it to text, means for analyzing video data to detect slide content and pointer movement, means for identifying speakers using facial recognition technology, means for integrating audio and video data to generate minutes, means for saving the generated minutes and making them searchable, means for extracting audio data from video, means for transmitting the extracted audio data to a voice recognition system and converting it to text data, and means for recording the content of lectures and question times. This allows the content of lectures and meetings to be automatically and accurately recorded, making it possible to easily search and reuse the information later.

[0469] "Audio data" refers to audio data recorded during a conference or lecture.

[0470] "Text data" is character string data converted from voice data using voice recognition technology.

[0471] "Video data" refers to video frame data recorded during a conference or lecture.

[0472] "Slide content" refers to information on presentation materials used during lectures and meetings.

[0473] "Pointer movement" refers to the movement trajectory of the cursor or pointer indicated on the screen during a presentation.

[0474] "Facial recognition technology" is a technology that identifies the face of a specific person from video data.

[0475] "Speaker" refers to the person who made a statement at a meeting or lecture.

[0476] "Integration" means combining individual audio data, video data, and speaker information into a single minutes.

[0477] "Minutes" are documents that record the progress of a meeting or lecture, what was said, slide information, etc.

[0478] "Searchable" means that stored information can be easily found based on specified criteria.

[0479] A "video" is a data file in which video and audio are recorded together.

[0480] A "voice recognition system" is a system that analyzes voice data and converts it into text data.

[0481] "Lecture content" refers to a series of statements, slides, and content intended for educational or explanatory purposes.

[0482] "Question windows" refer to specific time slots during lectures or meetings when questions arise.

[0483] This invention is a system that analyzes audio and video data from lectures and seminars and automatically generates detailed minutes. This system is realized using hardware such as a server, smartphone, and head-mounted display, and software such as a voice recognition module, video analysis module, and facial recognition technology.

[0484] System Overview

[0485] The system starts by a user uploading a video of a lecture or meeting to the server, which receives the video and performs the following steps:

[0486] 1. Extract audio data from video.

[0487] 2. The extracted voice data is sent to a voice recognition system and converted into text data.

[0488] 3. Analyze the video data to detect the slide content and pointer movement.

[0489] 4. Use facial recognition technology to identify speakers.

[0490] 5. Integrate these data to generate detailed meeting minutes.

[0491] 6. The generated minutes are saved on the server and made searchable.

[0492] Hardware and software used

[0493] Hardware:

[0494] Smartphones and head-mounted displays: Used to record and upload videos.

[0495] Server: Used to analyze data and generate transcripts.

[0496] software:

[0497] Python: Implementation of the main program and use of libraries.

[0498] cv2 (OpenCV): Video data analysis and frame segmentation.

[0499] speech_recognition: Converting speech data to text.

[0500] face_recognition: Application of facial recognition technology.

[0501] moviepy: Extracting audio data from videos.

[0502] transformers (Hugging Face): Speech recognition using the Wav2Vec2 model.

[0503] Processing flow

[0504] 1. Extracting audio data:

[0505] The server extracts the audio data from the uploaded video, which it does using the moviepy library.

[0506] 2. Speech Recognition:

[0507] The extracted audio data is converted into text data by a speech recognition system using the speech_recognition library. At this stage, the lecture content and speech content are obtained as text information.

[0508] 3. Video Data Analysis:

[0509] The server splits the video data into frames, detects the slide content and pointer movement using the OpenCV library, and identifies the speaker by recognizing the faces in each frame using the face_recognition library.

[0510] 4. Integration and Transcription:

[0511] The speech recognition results, video analysis results, and speaker information obtained at each step are integrated to generate minutes in the specified format. For example, the minutes will include information such as "Speaker A spoke about important point A while explaining slide 3."

[0512] 5. Archive and search meeting minutes:

[0513] The generated minutes are saved on the server for users to view and search later. Users can easily search for minutes based on "the content of a specific speaker" or "a specific slide."

[0514] Specific examples

[0515] For example, when a user uploads a video of an online lecture to the server, the system automatically analyzes the audio and video and generates detailed minutes that record the lecture content and the time periods when questions were asked. These minutes include detailed information such as, "Speaker A began explaining slide 1 immediately after the lecture began."

[0516] Prompt Sentence Examples

[0517] "Please convert the following audio data into text. Link to audio data: {URL of audio data}. Please describe the audio in chronological order and include speaker information if available."

[0518] As described above, the present invention makes it possible to automatically record the contents of lectures and meetings, and to easily search and reuse them later.

[0519] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0520] Step 1:

[0521] A user uploads video files of lectures or meetings to the server. The user selects the video files using their device and sends them to a specific upload destination on the server. The server stores the received video files and prepares them for further processing.

[0522] Input: Video file uploaded by user.

[0523] Output: Video file saved on the server.

[0524] Step 2:

[0525] The server extracts the audio data from the stored video file. For this process, it uses the moviepy library. The server reads the video file, separates the audio track, and saves it as a new audio file.

[0526] Input: Video file.

[0527] Output: Audio file.

[0528] Step 3:

[0529] The server sends the extracted audio file to a speech recognition module and converts it into text data. This process uses the speech_recognition library. The server reads the audio file and passes it through a speech recognition model to obtain text information.

[0530] Input: Audio file.

[0531] Output: Text data.

[0532] Step 4:

[0533] The server splits the video data from the video file into frames and analyzes them. This process uses the OpenCV library. The server splits the video file into frames and identifies the slide content, pointer movement, and speaker for each frame.

[0534] Input: Video file.

[0535] Output: Video data for each frame, analysis results (slide content, pointer movement, speaker information).

[0536] Step 5:

[0537] The server identifies the speaker based on the facial information acquired for each frame. Using the face_recognition library, the server compares the facial data in each frame with known facial data to identify the speaker.

[0538] Input: Face data for each frame.

[0539] Output: Identified speaker information.

[0540] Step 6:

[0541] The server then combines the acquired text data, slide content, pointer movement, and speaker information to generate detailed minutes. By combining each piece of data, the minutes are compiled in a format that users can easily review later.

[0542] Input: text data, slide content, pointer movement, speaker information.

[0543] Output: Detailed meeting transcript.

[0544] Step 7:

[0545] The server stores the generated minutes and provides them to users in a searchable format. The minutes are stored in a database and indexed so that users can easily search for specific criteria.

[0546] Input: Detailed minutes.

[0547] Output: Saved, searchable meeting minutes.

[0548] By using the above processing steps, the user can automatically record the contents of a lecture or conference, and later easily check the overall progress or the content of specific comments.

[0549] 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.

[0550] This invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates minutes of the meeting with an emotion engine that recognizes the emotions of users. This system processes audio and video data and utilizes the emotion engine to record the emotions of participants and reflect them in the minutes, making it possible to comprehensively understand the detailed situation of the meeting.

[0551] System Program Overview

[0552] The system begins by a user uploading a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers and an emotion engine to recognize participants' emotions. This data is then integrated to generate accurate minutes, which are then stored on the server and made available in a searchable format.

[0553] Audio data processing

[0554] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[0555] Video data processing

[0556] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[0557] Identifying the speaker

[0558] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[0559] Emotion recognition

[0560] Using the emotion engine, a distinctive feature of the present invention, the server recognizes participants' emotions from the video and audio data. The emotion engine uses techniques such as facial expression analysis and audio tone analysis to identify the emotional state (e.g., joy, anger, sadness, etc.) of each speaker. The recognized emotion information is saved as text data.

[0561] Generate meeting minutes

[0562] The server combines the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes in a specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[0563] Providing generated minutes

[0564] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[0565] Specific examples

[0566] For example, a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify the speaker, and an emotion engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Yamada spoke about important point A while explaining slide 3, and at the same time, Yamada seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[0567] This system not only records the details of the meeting, but also grasps the emotional state of the participants, allowing for deeper understanding and analysis.

[0568] The processing flow will be explained below.

[0569] Step 1:

[0570] The user uploads the video file of the conference from the terminal to the server.

[0571] Step 2:

[0572] The server receives the uploaded video file and stores it in a temporary storage area.

[0573] Step 3:

[0574] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[0575] Step 4:

[0576] The server uses a speech recognition module (e.g., Google Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[0577] Step 5:

[0578] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[0579] Step 6:

[0580] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[0581] Step 7:

[0582] The server uses a facial recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected facial data is matched with the audio data.

[0583] Step 8:

[0584] The server uses an emotion engine to recognize participants' emotions from video and audio data. The emotion engine analyzes facial expressions and audio tones, and stores the emotional state as text data.

[0585] Step 9:

[0586] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate a draft of the minutes, which includes text data for each statement, slide content, pointer movement, speaker information, and emotion information.

[0587] Step 10:

[0588] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[0589] Step 11:

[0590] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[0591] Example 2

[0592] 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."

[0593] Conventional meeting recording systems require a great deal of time and effort to process audio and video data separately and generate minutes manually. Furthermore, there is no way to grasp the emotional state of participants, making it difficult to understand the detailed situation of the meeting or the emotions of the participants. Furthermore, the generated minutes are not easily searchable or categorized, making them difficult to reference later.

[0594] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for extracting audio data and converting it into text, means for analyzing video data to detect visual information and indicator movements, means for identifying the caller using biometric recognition technology, means for recognizing an emotional state from the audio data and video data using a recognition engine, means for generating minutes by integrating the audio data, video data, caller information, and emotional information, and means for saving the generated minutes and making them categorizable. This enables audio and video to be processed in an integrated manner, automatically generating minutes that also include the caller's emotional state, and enabling efficient search and reference.

[0595] "Audio data" refers to recordings of sounds collected during meetings, conversations, etc.

[0596] "Text" is character information generated by processing audio data.

[0597] "Video data" refers to video information containing visual content such as meetings and presentations.

[0598] "Visual information" refers to information such as slide content and screen display analyzed from video data.

[0599] "Pointer movement" refers to interactive elements such as pointer or hand movements detected within the video data.

[0600] "Biometric recognition technology" is a technology that identifies individuals based on video data, such as facial recognition.

[0601] A "caller" is a person participating in a meeting or conversation.

[0602] A "cognitive engine" is a technology that analyzes audio and video data to recognize emotional states and other perceptual information.

[0603] "Emotional state" is information that indicates the psychological reactions and emotions (joy, anger, sadness, etc.) of the speaker or participants.

[0604] "Integration" is the process of combining various data to create a single, continuous record.

[0605] A "minutes" is a document that summarizes the contents of a meeting, what was said, the identity of the speakers, and their emotional state.

[0606] "Classifiable" means that the generated minutes are organized according to specific criteria or categories so that they can be easily searched and referenced.

[0607] A "system" is a structure having a series of integrated functions that includes all of the above means.

[0608] The present invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates meeting minutes with a recognition engine that recognizes user emotions. An embodiment of the present invention will be described in detail below.

[0609] The system starts by having a user upload a video of a meeting to the server, which receives the video and processes it in the following steps:

[0610] First, the server uses a library such as "FFmpeg" to extract audio data from the video. The extracted audio data is sent to a "speech recognition module" (general name) and converted into text data. For this purpose, a "speech recognition API" (Google Cloud Speech-to-Text API) or similar is used. This text data is then used to record what the participants are saying.

[0611] Next, the server divides the video into frames and processes the video data. A "video analysis module" (OpenCV) is used to divide the frames. The video analysis module analyzes each frame and detects the contents of the slides and the movement of the pointer. This information is used to record the details of the materials used in the meeting. In addition, "biometric recognition technology" (facial recognition technology) is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker.

[0612] An important feature is that the server uses a recognition engine, which includes technologies such as facial expression analysis and voice tone analysis (IBM Watson Tone Analyzer). Using these technologies, the server recognizes the emotional state (e.g., joy, anger, sadness, etc.) of each speaker from the video and voice data. The recognized emotional information is saved as text data.

[0613] Once this data is collected, the server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition, and generates minutes in the specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[0614] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[0615] As a concrete example, imagine a scenario where a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text using a "speech recognition API." Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Biometric recognition technology is used to identify the speaker, and a recognition engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Person A spoke about important point A while explaining slide 3, and at the same time, Person A seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[0616] An example prompt might be, "Upload a video of a project status meeting and automatically generate minutes. Please also record what Person A said and their emotions at the time."

[0617] The above is a specific description of the embodiment of the present invention. The present invention not only records the details of the meeting but also grasps the emotional states of the participants, making it possible to comprehensively understand and analyze the detailed situation of the meeting.

[0618] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0619] Step 1:

[0620] A user uploads a video of the meeting.

[0621] The user selects the video file of the "Project Progress Meeting" from their device and sends it to the server through the system's upload interface. As input, the video file of the meeting is passed from the device to the server. As output, the video file is saved on the server.

[0622] Step 2:

[0623] The server extracts the voice data and sends it to a voice recognition module.

[0624] The server analyzes the uploaded video file and extracts the audio data using a library such as FFmpeg. This audio data is then sent to a speech recognition API. The audio data extracted from the video file is used as input. The audio data is converted to text as output.

[0625] Step 3:

[0626] The server divides the video data into frames and processes them in a video analysis module.

[0627] The server splits the video data into frames using OpenCV. For each frame, it uses OCR technology to recognize the slide content and analyzes the pointer, hand, and other indicator movements. The video data is used as input. Data about the slide content and indicator movements is generated as output.

[0628] Step 4:

[0629] The server uses facial recognition technology to identify the speaker.

[0630] The server detects facial images in each frame of video data and identifies the speaker using facial recognition technology such as Amazon Rekognition. The detected facial images are then matched with the audio data. The facial images and audio data are used as input. The output is data that identifies the speaker.

[0631] Step 5:

[0632] The server uses an emotion engine to recognize the emotions of the participants.

[0633] The server uses an emotion engine to analyze the video and audio data and recognize the speaker's emotional state. This process uses facial expression analysis and tone analysis techniques. The video and audio data are used as input. The output is an emotion recognition result (e.g., joy, anger, sadness).

[0634] Step 6:

[0635] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes.

[0636] The server integrates the data collected so far and generates minutes in the specified format. These minutes include detailed records of the progress of the meeting, the contents of the materials, and the remarks and emotions of the speakers. Various analysis and recognition results are used as input. The integrated minutes data is generated as output.

[0637] Step 7:

[0638] The server provides the generated minutes.

[0639] The generated minutes are stored on a server and provided in a format that users can access later for searching and viewing. Specifically, users can use the search interface to search for "emotional reactions to a specific slide in a meeting," "the content of a specific speaker's remarks and their emotions at the time," and so on. The user's search query is used as input, and the relevant minutes data is provided as output.

[0640] The above is the specific processing flow of the system program divided into specific processing steps, and the details of each process accompanied by input and output.

[0641] (Application example 2)

[0642] 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."

[0643] Conventional meeting recording systems convert audio data into text and analyze video data, but lack the functionality to monitor participants' emotions and work environment in real time. This makes it difficult to understand the details of the meeting content or to properly grasp the fatigue and stress levels of workers. Furthermore, there is a need for a system that can respond in real time to changes in the work environment.

[0644] 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 extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movement, means for identifying the speaker using face recognition technology, means for recognizing the emotions of participants using emotion recognition technology, means for integrating audio data and video data to generate minutes, means for saving the generated minutes and making them searchable, and means for monitoring the work environment in real time and detecting the fatigue and stress of workers. This makes it possible to not only record detailed information about meetings and work content, but also to monitor the emotions of participants and the work environment in real time.

[0645] "Voice data" refers to voice information such as statements and instructions that occur during meetings or work.

[0646] "Means for converting to text" refers to the technology or equipment that converts voice data into text information.

[0647] "Video data" refers to video information captured by cameras or recording devices.

[0648] "Slide content" refers to the information on slides used in presentations or as meeting materials.

[0649] "Indicator movement" refers to the movement of a pointer, mouse cursor, or other device on the screen to indicate instructions or a point of interest.

[0650] "Facial recognition technology" refers to technology for recognizing and identifying specific individuals from video data.

[0651] "Means for identifying the speaker" refers to technologies and methods for determining who is speaking based on audio and video data.

[0652] "Emotion recognition technology" refers to technology for identifying an individual's emotional state (joy, anger, sadness, etc.) from audio and video data.

[0653] "Means for recognizing emotions of participants" refers to a technique or method for determining the emotions of participants in a meeting or workplace using emotion recognition technology.

[0654] "Means for generating minutes" refers to technologies and methods that integrate audio data, video data, emotional information, etc. to automatically generate documents that record the content of meetings and work.

[0655] "Means for storing and making searchable minutes" refers to the technology and methods for electronically storing the generated minutes in a database so that they can be searched and viewed at a later date.

[0656] "Means for monitoring the work environment in real time" refers to technologies and methods for instantly monitoring the progress and environment of work in factories and workplaces.

[0657] "Means for detecting worker fatigue and stress" refers to technologies and methods for assessing the health and mental state of workers using emotion recognition technology or other sensor technology.

[0658] The present invention relates to a work monitoring and quality control system using factory robots. The system includes means for extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movements, means for identifying speakers using facial recognition technology, means for recognizing participants' emotions using emotion recognition technology, means for integrating audio data and video data to generate minutes of meetings, means for storing the generated minutes and making them searchable, and means for monitoring the work environment in real time to detect worker fatigue and stress.

[0659] System Program Overview

[0660] The server collects video and audio data from within the factory using high-resolution cameras and microphones. The collected audio data is converted into text using a speech recognition module (e.g., IBM Watson Speech to Text). The video data is divided into frames and analyzed using OpenCV. The speaker is identified using slide content, indicator movements, and facial recognition technology.

[0661] Using emotion recognition technology (e.g., Affectiva SDK), the emotions of workers are recognized from video and audio data. This allows participants' emotional information to be saved as text data. The results of audio recognition, video analysis, speaker identification, and emotion recognition are integrated to generate minutes in the specified format.

[0662] Hardware and Software Use

[0663] The system uses a high-resolution camera, microphone, high-performance CPU and GPU, and utilizes TensorFlow (machine learning library), OpenCV (image processing library), IBM Watson Speech to Text (voice recognition), and Affectiva SDK (emotion recognition) software.

[0664] Specific examples

[0665] For example, if a factory worker says, "I can't assemble this part properly," the voice recording is collected and converted into text. Emotion recognition technology recognizes that the worker is feeling "frustrated." This information is collected and analyzed as follows:

[0666] Audio data:

[0667] "This part doesn't assemble properly" -> recognized_text: "This part doesn't assemble properly"

[0668] Emotional Data:

[0669] recognized_emotion: "frustration"

[0670] We also provide an example of input to a generative AI model using the following prompt sentence:

[0671] "Generate a monitoring report for the factory. Analyze the voice and emotions of the workers and incorporate them into the report."

[0672] This allows the system to record detailed factory working conditions as well as monitor workers' emotions and health in real time, allowing appropriate responses to be taken.

[0673] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0674] Step 1:

[0675] The server uses microphones to collect voice data from within the factory. This allows it to obtain voice information such as conversations and instructions within the factory. Specifically, the analog voice signals collected by the microphones are converted into digital data and input into a voice recognition module. The input is voice data, and the output is digitized voice data.

[0676] Step 2:

[0677] The server uses a speech recognition module (e.g., IBM Watson Speech to Text) to convert the collected voice data into text. The input is digital voice data, and the output is text data. Specifically, the voice signal is converted into acoustic features, which are then input into a model to generate a corresponding string of characters.

[0678] Step 3:

[0679] The server uses a high-resolution camera to collect video data within the factory, visually recording the progress and environment of work. The input is visual information and the output is video data. The video data collected by the camera is divided into frames and input into the image processing module.

[0680] Step 4:

[0681] The server divides the video data into frames and analyzes them using OpenCV. The input is the video data, and the output is analyzed data on the slide content and indicator movement. Specifically, image processing is performed on each frame to extract the characteristics of the slides and indicators.

[0682] Step 5:

[0683] The server uses facial recognition technology (e.g., OpenCV's facial recognition module) to identify speakers from video data. The input is video data, and the output is data on the identified speaker. Specifically, it detects facial features in the video frame and compares them with an existing database to identify individuals.

[0684] Step 6:

[0685] The server uses emotion recognition technology (e.g., Affectiva SDK) to recognize participants' emotions from the collected audio and video data. The input is audio and video data, and the output is emotional state data. Specifically, it performs voice tone analysis and facial expression analysis to identify emotional levels.

[0686] Step 7:

[0687] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to automatically generate minutes in a specified format. The input is a dataset of each analysis result, and the output is the integrated minutes. Specifically, each piece of data is organized chronologically and a document is generated in a specified format.

[0688] Step 8:

[0689] The server stores the generated minutes in a database and provides them in a format that can be searched later. The input is the generated minutes, and the output is a searchable database. Specifically, the minutes data is indexed and stored in the database, making it searchable later using queries.

[0690] Step 9:

[0691] The server monitors the work environment in real time and detects the fatigue and stress levels of workers. The input is video, audio, and emotional data, and the output is data on the fatigue and stress levels of workers. Specifically, it analyzes continuously collected data and generates an alert if a certain threshold is exceeded.

[0692] In this way, it is possible to monitor participants' emotions and work environment in real time, along with detailed records of meetings and work content, and take appropriate measures.

[0693] 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.

[0694] 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.

[0695] 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.

[0696] [Third embodiment]

[0697] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0698] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0699] 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).

[0700] 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.

[0701] 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.

[0702] 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).

[0703] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0704] 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.

[0705] 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.

[0706] 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.

[0707] 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.

[0708] 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."

[0709] This invention is a system that analyzes the audio and video of a meeting and automatically generates detailed minutes. This system processes audio and video data, accurately capturing the content of materials and participants' remarks, and creates minutes.

[0710] System Program Overview

[0711] The system begins by having a user upload a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers. This data is then integrated to generate accurate minutes. The generated minutes are then stored on the server and made available in a searchable format.

[0712] Audio data processing

[0713] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[0714] Video data processing

[0715] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[0716] Identifying the speaker

[0717] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[0718] Generate meeting minutes

[0719] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in the specified format, which contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[0720] Providing generated minutes

[0721] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search for minutes based on "what a specific speaker said" or "a specific slide from the meeting."

[0722] Specific examples

[0723] For example, a user uploads a video of a "project progress meeting" to the server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify speakers, and this information is integrated to generate minutes. The generated minutes include details such as "Yamada spoke about important point A while explaining slide 3." These minutes are stored on the server and can be searched and viewed by users later.

[0724] This system allows detailed recording of meeting contents, allowing users to easily check and reuse the information later.

[0725] The processing flow will be explained below.

[0726] Step 1:

[0727] The user uploads the video file of the conference from the terminal to the server.

[0728] Step 2:

[0729] The server receives the uploaded video file and stores it in a temporary storage area.

[0730] Step 3:

[0731] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[0732] Step 4:

[0733] The server uses a speech recognition module (e.g., Google Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[0734] Step 5:

[0735] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[0736] Step 6:

[0737] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[0738] Step 7:

[0739] The server uses a face recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected face data is matched with the audio data.

[0740] Step 8:

[0741] The server combines the results of speech recognition, video analysis, and speaker identification to generate a draft of the minutes, which includes the text data of each statement, slide content, pointer movement, and speaker information.

[0742] Step 9:

[0743] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[0744] Step 10:

[0745] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[0746] Example 1

[0747] 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."

[0748] Accurately analyzing the audio and video of a meeting and automatically generating detailed minutes is a time-consuming and labor-intensive process. Furthermore, it is often difficult to accurately identify speakers and record slide content. Furthermore, it is difficult to quickly search through the generated minutes to find specific information.

[0749] 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.

[0750] In this invention, the server includes means for extracting audio data and converting it to text, means for dividing and analyzing video data by frame to detect slide content and pointer movement, means for identifying speakers using face recognition technology, means for generating minutes by integrating the results of the audio recognition, video analysis, and speaker identification, and means for saving the generated minutes and providing them in a searchable format, thereby enabling accurate recording of the meeting contents and rapid subsequent retrieval.

[0751] "Audio data" refers to digital data that is a recording of what is said during a meeting or the like.

[0752] "Video data" refers to digital data in the form of moving images captured during a conference.

[0753] A "voice recognition module" is a software or hardware component for converting voice data into text data.

[0754] A "frame" is a unit that refers to each still image that makes up video data.

[0755] A "video analysis module" is a software or hardware component that analyzes video data to detect slide content and pointer movement.

[0756] "Slide content" refers to the information on presentation slides displayed during a conference.

[0757] "Pointer movement" refers to the movement of the pointer or mouse on the slides during the meeting.

[0758] "Facial recognition technology" is a technology for identifying and recognizing individual faces from video data.

[0759] "Speaker" refers to a person who speaks during a meeting.

[0760] A "minutes" is a document that records in detail the progress of a meeting, what was said, the contents of slides, etc.

[0761] A "searchable format" is a data format that is indexed so that a user can later search for specific content.

[0762] The present invention is a system for automatically generating detailed meeting minutes by analyzing audio and video of a meeting. The system uses multiple hardware and software modules to process audio and video data and generate accurate meeting minutes.

[0763] Basic configuration

[0764] Audio data processing

[0765] The system begins operation when a user uploads a video file of a meeting to the server. The server first extracts audio data from the video file using a tool such as FFmpeg. This extracted audio data is then sent to a speech recognition module (e.g., Google Cloud Speech-to-Text API) and converted into text data. This text data is then used to record what was said by the meeting participants.

[0766] Video data processing

[0767] Next, the server splits the video file into frames using a tool such as FFmpeg. Each frame is then analyzed using a video analysis module such as OpenCV. This analysis detects the slide content and pointer movement.

[0768] Identifying the speaker

[0769] Facial recognition technology (e.g., Amazon Rekognition) is used to analyze the video data. The server uses the video data to detect the faces of each participant and identifies the speaker by matching them with the audio data. This makes it clear which participant made each comment.

[0770] Generate meeting minutes

[0771] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in a specified format (e.g., Markdown or PDF). These minutes contain detailed records of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[0772] Providing generated minutes

[0773] The generated minutes are saved on the server. The saved minutes are indexed using Elasticsearch and provided in a format that makes them easy for users to search later. Users can access the minutes using a web browser or a dedicated app, and search and view them based on "the content of a specific speaker" or "a specific slide."

[0774] Specific examples

[0775] For example, a user uploads a video file of a "project progress meeting" to the server. This video file can be in supported formats such as MP4 or AVI. The server first extracts the audio data and converts it to text using the Google Cloud Speech-to-Text API. Next, the server divides the video data into frames and analyzes the slide content and pointer movement using OpenCV. It identifies speakers using facial recognition technology (e.g., Amazon Rekognition), and combines this data to generate minutes. The generated minutes include details such as "Participant A spoke about an important point while explaining slide 3." The minutes are stored on the server and can be searched and viewed later by users.

[0776] Prompt Sentence Examples

[0777] I have uploaded a video of a "Project Status Meeting." Please generate minutes of this meeting. The minutes should include the content of each slide and what each speaker said.

[0778] This system automatically and accurately records detailed meeting content, allowing users to quickly search for and check the information they need later.

[0779] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0780] Step 1:

[0781] The user uploads the video file of the conference to the server. At this stage, the video file is sent to the server and saved in the specified directory. The input is the video file provided from the user's device, and the output is the video file saved on the server side.

[0782] Step 2:

[0783] The server receives the video file and extracts the audio data. Using a tool such as FFmpeg, it separates the audio track from the video file. The input is the video file saved in step 1, and the output is an audio file temporarily stored on the server.

[0784] Step 3:

[0785] The server sends the audio data to the speech recognition module. The audio file is converted into text data using a method such as the Google Cloud Speech-to-Text API. During this process, the speech recognition module is called by the server. The input is the audio file, and the output is text data.

[0786] Step 4:

[0787] The server receives the text data from the speech recognition module and stores it as the speech content. This allows each speech made during the meeting to be recorded in text format. The input is the text data of the speech recognition results, and the output is the storage of the text data as the speech content.

[0788] Step 5:

[0789] The server splits the video file into frames. FFmpeg is used to convert the video into frames, which are a series of still images. The input is the video file saved in step 1, and the output is an image file split into frames.

[0790] Step 6:

[0791] The server sends the image files divided into frames to a video analysis module for analysis. Using a video analysis module such as OpenCV, slide content and pointer movement are detected from each frame. The input is the frame image file, and the output is the analysis results: slide content and pointer movement information.

[0792] Step 7:

[0793] The server uses facial recognition technology to identify the speaker. Using Amazon Rekognition or similar, it detects the face of each participant from the video data and compares it with the audio data. The input is the frame image file and audio data, and the output is data identifying the speaker.

[0794] Step 8:

[0795] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes. The minutes are created in the specified format (Markdown, PDF, etc.). The input is the data obtained at each processing step (text data, analysis results, speaker information), and the output is the completed minutes file.

[0796] Step 9:

[0797] The generated minutes are stored on a server and provided in a searchable format. They are indexed using a search engine such as Elasticsearch, allowing users to easily access and search them later. The input is the generated minutes file, and the output is an indexed database.

[0798] Step 10:

[0799] Users search and view meeting minutes using a web browser or a dedicated app. The server returns indexed meeting minutes in response to the user's request. The input is the user's search query, and the output is the relevant part of the meeting minutes as a search result.

[0800] (Application example 1)

[0801] 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."

[0802] Conventional meeting minutes systems require manual processing of audio and video data to create minutes, requiring a great deal of time and effort. Similar problems existed in online lectures and seminars, making it difficult to accurately record the content of lectures and details of Q&A sessions. Furthermore, searching this data later required a great deal of effort, preventing sufficient reuse of information.

[0803] 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.

[0804] In this invention, the server includes means for extracting audio data and converting it to text, means for analyzing video data to detect slide content and pointer movement, means for identifying speakers using facial recognition technology, means for integrating audio and video data to generate minutes, means for saving the generated minutes and making them searchable, means for extracting audio data from video, means for transmitting the extracted audio data to a voice recognition system and converting it to text data, and means for recording the content of lectures and question times. This allows the content of lectures and meetings to be automatically and accurately recorded, making it possible to easily search and reuse the information later.

[0805] "Audio data" refers to audio data recorded during a conference or lecture.

[0806] "Text data" is character string data converted from voice data using voice recognition technology.

[0807] "Video data" refers to video frame data recorded during a conference or lecture.

[0808] "Slide content" refers to information on presentation materials used during lectures and meetings.

[0809] "Pointer movement" refers to the movement trajectory of the cursor or pointer indicated on the screen during a presentation.

[0810] "Facial recognition technology" is a technology that identifies the face of a specific person from video data.

[0811] "Speaker" refers to the person who made a statement at a meeting or lecture.

[0812] "Integration" means combining individual audio data, video data, and speaker information into a single minutes.

[0813] "Minutes" are documents that record the progress of a meeting or lecture, what was said, slide information, etc.

[0814] "Searchable" means that stored information can be easily found based on specified criteria.

[0815] A "video" is a data file in which video and audio are recorded together.

[0816] A "voice recognition system" is a system that analyzes voice data and converts it into text data.

[0817] "Lecture content" refers to a series of statements, slides, and content intended for educational or explanatory purposes.

[0818] "Question windows" refer to specific time slots during lectures or meetings when questions arise.

[0819] This invention is a system that analyzes audio and video data from lectures and seminars and automatically generates detailed minutes. This system is realized using hardware such as a server, smartphone, and head-mounted display, and software such as a voice recognition module, video analysis module, and facial recognition technology.

[0820] System Overview

[0821] The system starts by a user uploading a video of a lecture or meeting to the server, which receives the video and performs the following steps:

[0822] 1. Extract audio data from video.

[0823] 2. The extracted voice data is sent to a voice recognition system and converted into text data.

[0824] 3. Analyze the video data to detect the slide content and pointer movement.

[0825] 4. Use facial recognition technology to identify speakers.

[0826] 5. Integrate these data to generate detailed meeting minutes.

[0827] 6. The generated minutes are saved on the server and made searchable.

[0828] Hardware and software used

[0829] Hardware:

[0830] Smartphones and head-mounted displays: Used to record and upload videos.

[0831] Server: Used to analyze data and generate transcripts.

[0832] software:

[0833] Python: Implementation of the main program and use of libraries.

[0834] cv2 (OpenCV): Video data analysis and frame segmentation.

[0835] speech_recognition: Converting speech data to text.

[0836] face_recognition: Application of facial recognition technology.

[0837] moviepy: Extracting audio data from videos.

[0838] transformers (Hugging Face): Speech recognition using the Wav2Vec2 model.

[0839] Processing flow

[0840] 1. Extracting audio data:

[0841] The server extracts the audio data from the uploaded video, which it does using the moviepy library.

[0842] 2. Speech Recognition:

[0843] The extracted audio data is converted into text data by a speech recognition system using the speech_recognition library. At this stage, the lecture content and speech content are obtained as text information.

[0844] 3. Video Data Analysis:

[0845] The server splits the video data into frames, detects the slide content and pointer movement using the OpenCV library, and identifies the speaker by recognizing the faces in each frame using the face_recognition library.

[0846] 4. Integration and Transcription:

[0847] The speech recognition results, video analysis results, and speaker information obtained at each step are integrated to generate minutes in the specified format. For example, the minutes will include information such as "Speaker A spoke about important point A while explaining slide 3."

[0848] 5. Archive and search meeting minutes:

[0849] The generated minutes are saved on the server for users to view and search later. Users can easily search for minutes based on "the content of a specific speaker" or "a specific slide."

[0850] Specific examples

[0851] For example, when a user uploads a video of an online lecture to the server, the system automatically analyzes the audio and video and generates detailed minutes that record the lecture content and the time periods when questions were asked. These minutes include detailed information such as, "Speaker A began explaining slide 1 immediately after the lecture began."

[0852] Prompt Sentence Examples

[0853] "Please convert the following audio data into text. Link to audio data: {URL of audio data}. Please describe the audio in chronological order and include speaker information if available."

[0854] As described above, the present invention makes it possible to automatically record the contents of lectures and meetings, and to easily search and reuse them later.

[0855] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0856] Step 1:

[0857] A user uploads video files of lectures or meetings to the server. The user selects the video files using their device and sends them to a specific upload destination on the server. The server stores the received video files and prepares them for further processing.

[0858] Input: Video file uploaded by user.

[0859] Output: Video file saved on the server.

[0860] Step 2:

[0861] The server extracts the audio data from the stored video file. For this process, it uses the moviepy library. The server reads the video file, separates the audio track, and saves it as a new audio file.

[0862] Input: Video file.

[0863] Output: Audio file.

[0864] Step 3:

[0865] The server sends the extracted audio file to a speech recognition module and converts it into text data. This process uses the speech_recognition library. The server reads the audio file and passes it through a speech recognition model to obtain text information.

[0866] Input: Audio file.

[0867] Output: Text data.

[0868] Step 4:

[0869] The server splits the video data from the video file into frames and analyzes them. This process uses the OpenCV library. The server splits the video file into frames and identifies the slide content, pointer movement, and speaker for each frame.

[0870] Input: Video file.

[0871] Output: Video data for each frame, analysis results (slide content, pointer movement, speaker information).

[0872] Step 5:

[0873] The server identifies the speaker based on the facial information acquired for each frame. Using the face_recognition library, the server compares the facial data in each frame with known facial data to identify the speaker.

[0874] Input: Face data for each frame.

[0875] Output: Identified speaker information.

[0876] Step 6:

[0877] The server then combines the acquired text data, slide content, pointer movement, and speaker information to generate detailed minutes. By combining each piece of data, the minutes are compiled in a format that users can easily review later.

[0878] Input: text data, slide content, pointer movement, speaker information.

[0879] Output: Detailed meeting transcript.

[0880] Step 7:

[0881] The server stores the generated minutes and provides them to users in a searchable format. The minutes are stored in a database and indexed so that users can easily search for specific criteria.

[0882] Input: Detailed minutes.

[0883] Output: Saved, searchable meeting minutes.

[0884] By using the above processing steps, the user can automatically record the contents of a lecture or conference, and later easily check the overall progress or the content of specific comments.

[0885] 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.

[0886] This invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates minutes of the meeting with an emotion engine that recognizes the emotions of users. This system processes audio and video data and utilizes the emotion engine to record the emotions of participants and reflect them in the minutes, making it possible to comprehensively understand the detailed situation of the meeting.

[0887] System Program Overview

[0888] The system begins by a user uploading a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers and an emotion engine to recognize participants' emotions. This data is then integrated to generate accurate minutes, which are then stored on the server and made available in a searchable format.

[0889] Audio data processing

[0890] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[0891] Video data processing

[0892] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[0893] Identifying the speaker

[0894] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[0895] Emotion recognition

[0896] Using the emotion engine, a distinctive feature of the present invention, the server recognizes participants' emotions from the video and audio data. The emotion engine uses techniques such as facial expression analysis and audio tone analysis to identify the emotional state (e.g., joy, anger, sadness, etc.) of each speaker. The recognized emotion information is saved as text data.

[0897] Generate meeting minutes

[0898] The server combines the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes in a specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[0899] Providing generated minutes

[0900] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[0901] Specific examples

[0902] For example, a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify the speaker, and an emotion engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Yamada spoke about important point A while explaining slide 3, and at the same time, Yamada seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[0903] This system not only records the details of the meeting, but also grasps the emotional state of the participants, allowing for deeper understanding and analysis.

[0904] The processing flow will be explained below.

[0905] Step 1:

[0906] The user uploads the video file of the conference from the terminal to the server.

[0907] Step 2:

[0908] The server receives the uploaded video file and stores it in a temporary storage area.

[0909] Step 3:

[0910] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[0911] Step 4:

[0912] The server uses a speech recognition module (e.g., Google Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[0913] Step 5:

[0914] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[0915] Step 6:

[0916] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[0917] Step 7:

[0918] The server uses a facial recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected facial data is matched with the audio data.

[0919] Step 8:

[0920] The server uses an emotion engine to recognize participants' emotions from video and audio data. The emotion engine analyzes facial expressions and audio tones, and stores the emotional state as text data.

[0921] Step 9:

[0922] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate a draft of the minutes, which includes text data for each statement, slide content, pointer movement, speaker information, and emotion information.

[0923] Step 10:

[0924] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[0925] Step 11:

[0926] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[0927] Example 2

[0928] 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."

[0929] Conventional meeting recording systems require a great deal of time and effort to process audio and video data separately and generate minutes manually. Furthermore, there is no way to grasp the emotional state of participants, making it difficult to understand the detailed situation of the meeting or the emotions of the participants. Furthermore, the generated minutes are not easily searchable or categorized, making them difficult to reference later.

[0930] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for extracting audio data and converting it into text, means for analyzing video data to detect visual information and indicator movements, means for identifying the caller using biometric recognition technology, means for recognizing an emotional state from the audio data and video data using a recognition engine, means for generating minutes by integrating the audio data, video data, caller information, and emotional information, and means for saving the generated minutes and making them categorizable. This enables audio and video to be processed in an integrated manner, automatically generating minutes that also include the caller's emotional state, and enabling efficient search and reference.

[0931] "Audio data" refers to recordings of sounds collected during meetings, conversations, etc.

[0932] "Text" is character information generated by processing audio data.

[0933] "Video data" refers to video information containing visual content such as meetings and presentations.

[0934] "Visual information" refers to information such as slide content and screen display analyzed from video data.

[0935] "Pointer movement" refers to interactive elements such as pointer or hand movements detected within the video data.

[0936] "Biometric recognition technology" is a technology that identifies individuals based on video data, such as facial recognition.

[0937] A "caller" is a person participating in a meeting or conversation.

[0938] A "cognitive engine" is a technology that analyzes audio and video data to recognize emotional states and other perceptual information.

[0939] "Emotional state" is information that indicates the psychological reactions and emotions (joy, anger, sadness, etc.) of the speaker or participants.

[0940] "Integration" is the process of combining various data to create a single, continuous record.

[0941] A "minutes" is a document that summarizes the contents of a meeting, what was said, the identity of the speakers, and their emotional state.

[0942] "Classifiable" means that the generated minutes are organized according to specific criteria or categories so that they can be easily searched and referenced.

[0943] A "system" is a structure having a series of integrated functions that includes all of the above means.

[0944] The present invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates meeting minutes with a recognition engine that recognizes user emotions. An embodiment of the present invention will be described in detail below.

[0945] The system starts by having a user upload a video of a meeting to the server, which receives the video and processes it in the following steps:

[0946] First, the server uses a library such as "FFmpeg" to extract audio data from the video. The extracted audio data is sent to a "speech recognition module" (general name) and converted into text data. For this purpose, a "speech recognition API" (Google Cloud Speech-to-Text API) or similar is used. This text data is then used to record what the participants are saying.

[0947] Next, the server divides the video into frames and processes the video data. A "video analysis module" (OpenCV) is used to divide the frames. The video analysis module analyzes each frame and detects the contents of the slides and the movement of the pointer. This information is used to record the details of the materials used in the meeting. In addition, "biometric recognition technology" (facial recognition technology) is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker.

[0948] An important feature is that the server uses a recognition engine, which includes technologies such as facial expression analysis and voice tone analysis (IBM Watson Tone Analyzer). Using these technologies, the server recognizes the emotional state (e.g., joy, anger, sadness, etc.) of each speaker from the video and voice data. The recognized emotional information is saved as text data.

[0949] Once this data is collected, the server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition, and generates minutes in the specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[0950] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[0951] As a concrete example, imagine a scenario where a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text using a "speech recognition API." Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Biometric recognition technology is used to identify the speaker, and a recognition engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Person A spoke about important point A while explaining slide 3, and at the same time, Person A seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[0952] An example prompt might be, "Upload a video of a project status meeting and automatically generate minutes. Please also record what Person A said and their emotions at the time."

[0953] The above is a specific description of the embodiment of the present invention. The present invention not only records the details of the meeting but also grasps the emotional states of the participants, making it possible to comprehensively understand and analyze the detailed situation of the meeting.

[0954] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0955] Step 1:

[0956] A user uploads a video of the meeting.

[0957] The user selects the video file of the "Project Progress Meeting" from their device and sends it to the server through the system's upload interface. As input, the video file of the meeting is passed from the device to the server. As output, the video file is saved on the server.

[0958] Step 2:

[0959] The server extracts the voice data and sends it to a voice recognition module.

[0960] The server analyzes the uploaded video file and extracts the audio data using a library such as FFmpeg. This audio data is then sent to a speech recognition API. The audio data extracted from the video file is used as input. The audio data is converted to text as output.

[0961] Step 3:

[0962] The server divides the video data into frames and processes them in a video analysis module.

[0963] The server splits the video data into frames using OpenCV. For each frame, it uses OCR technology to recognize the slide content and analyzes the pointer, hand, and other indicator movements. The video data is used as input. Data about the slide content and indicator movements is generated as output.

[0964] Step 4:

[0965] The server uses facial recognition technology to identify the speaker.

[0966] The server detects facial images in each frame of video data and identifies the speaker using facial recognition technology such as Amazon Rekognition. The detected facial images are then matched with the audio data. The facial images and audio data are used as input. The output is data that identifies the speaker.

[0967] Step 5:

[0968] The server uses an emotion engine to recognize the emotions of the participants.

[0969] The server uses an emotion engine to analyze the video and audio data and recognize the speaker's emotional state. This process uses facial expression analysis and tone analysis techniques. The video and audio data are used as input. The output is an emotion recognition result (e.g., joy, anger, sadness).

[0970] Step 6:

[0971] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes.

[0972] The server integrates the data collected so far and generates minutes in the specified format. These minutes include detailed records of the progress of the meeting, the contents of the materials, and the remarks and emotions of the speakers. Various analysis and recognition results are used as input. The integrated minutes data is generated as output.

[0973] Step 7:

[0974] The server provides the generated minutes.

[0975] The generated minutes are stored on a server and provided in a format that users can access later for searching and viewing. Specifically, users can use the search interface to search for "emotional reactions to a specific slide in a meeting," "the content of a specific speaker's remarks and their emotions at the time," and so on. The user's search query is used as input, and the relevant minutes data is provided as output.

[0976] The above is the specific processing flow of the system program divided into specific processing steps, and the details of each process accompanied by input and output.

[0977] (Application example 2)

[0978] 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."

[0979] Conventional meeting recording systems convert audio data into text and analyze video data, but lack the functionality to monitor participants' emotions and work environment in real time. This makes it difficult to understand the details of the meeting content or to properly grasp the fatigue and stress levels of workers. Furthermore, there is a need for a system that can respond in real time to changes in the work environment.

[0980] 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 extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movement, means for identifying the speaker using face recognition technology, means for recognizing the emotions of participants using emotion recognition technology, means for integrating audio data and video data to generate minutes, means for saving the generated minutes and making them searchable, and means for monitoring the work environment in real time and detecting the fatigue and stress of workers. This makes it possible to not only record detailed information about meetings and work content, but also to monitor the emotions of participants and the work environment in real time.

[0981] "Voice data" refers to voice information such as statements and instructions that occur during meetings or work.

[0982] "Means for converting to text" refers to the technology or equipment that converts voice data into text information.

[0983] "Video data" refers to video information captured by cameras or recording devices.

[0984] "Slide content" refers to the information on slides used in presentations or as meeting materials.

[0985] "Indicator movement" refers to the movement of a pointer, mouse cursor, or other device on the screen to indicate instructions or a point of interest.

[0986] "Facial recognition technology" refers to technology for recognizing and identifying specific individuals from video data.

[0987] "Means for identifying the speaker" refers to technologies and methods for determining who is speaking based on audio and video data.

[0988] "Emotion recognition technology" refers to technology for identifying an individual's emotional state (joy, anger, sadness, etc.) from audio and video data.

[0989] "Means for recognizing emotions of participants" refers to a technique or method for determining the emotions of participants in a meeting or workplace using emotion recognition technology.

[0990] "Means for generating minutes" refers to technologies and methods that integrate audio data, video data, emotional information, etc. to automatically generate documents that record the content of meetings and work.

[0991] "Means for storing and making searchable minutes" refers to the technology and methods for electronically storing the generated minutes in a database so that they can be searched and viewed at a later date.

[0992] "Means for monitoring the work environment in real time" refers to technologies and methods for instantly monitoring the progress and environment of work in factories and workplaces.

[0993] "Means for detecting worker fatigue and stress" refers to technologies and methods for assessing the health and mental state of workers using emotion recognition technology or other sensor technology.

[0994] The present invention relates to a work monitoring and quality control system using factory robots. The system includes means for extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movements, means for identifying speakers using facial recognition technology, means for recognizing participants' emotions using emotion recognition technology, means for integrating audio data and video data to generate minutes of meetings, means for storing the generated minutes and making them searchable, and means for monitoring the work environment in real time to detect worker fatigue and stress.

[0995] System Program Overview

[0996] The server collects video and audio data from within the factory using high-resolution cameras and microphones. The collected audio data is converted into text using a speech recognition module (e.g., IBM Watson Speech to Text). The video data is divided into frames and analyzed using OpenCV. The speaker is identified using slide content, indicator movements, and facial recognition technology.

[0997] Using emotion recognition technology (e.g., Affectiva SDK), the emotions of workers are recognized from video and audio data. This allows participants' emotional information to be saved as text data. The results of audio recognition, video analysis, speaker identification, and emotion recognition are integrated to generate minutes in the specified format.

[0998] Hardware and Software Use

[0999] The system uses a high-resolution camera, microphone, high-performance CPU and GPU, and utilizes TensorFlow (machine learning library), OpenCV (image processing library), IBM Watson Speech to Text (voice recognition), and Affectiva SDK (emotion recognition) software.

[1000] Specific examples

[1001] For example, if a factory worker says, "I can't assemble this part properly," the voice recording is collected and converted into text. Emotion recognition technology recognizes that the worker is feeling "frustrated." This information is collected and analyzed as follows:

[1002] Audio data:

[1003] "This part doesn't assemble properly" -> recognized_text: "This part doesn't assemble properly"

[1004] Emotional Data:

[1005] recognized_emotion: "frustration"

[1006] We also provide an example of input to a generative AI model using the following prompt sentence:

[1007] "Generate a monitoring report for the factory. Analyze the voice and emotions of the workers and incorporate them into the report."

[1008] This allows the system to record detailed factory working conditions as well as monitor workers' emotions and health in real time, allowing appropriate responses to be taken.

[1009] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1010] Step 1:

[1011] The server uses microphones to collect voice data from within the factory. This allows it to obtain voice information such as conversations and instructions within the factory. Specifically, the analog voice signals collected by the microphones are converted into digital data and input into a voice recognition module. The input is voice data, and the output is digitized voice data.

[1012] Step 2:

[1013] The server uses a speech recognition module (e.g., IBM Watson Speech to Text) to convert the collected voice data into text. The input is digital voice data, and the output is text data. Specifically, the voice signal is converted into acoustic features, which are then input into a model to generate a corresponding string of characters.

[1014] Step 3:

[1015] The server uses a high-resolution camera to collect video data within the factory, visually recording the progress and environment of work. The input is visual information and the output is video data. The video data collected by the camera is divided into frames and input into the image processing module.

[1016] Step 4:

[1017] The server divides the video data into frames and analyzes them using OpenCV. The input is the video data, and the output is analyzed data on the slide content and indicator movement. Specifically, image processing is performed on each frame to extract the characteristics of the slides and indicators.

[1018] Step 5:

[1019] The server uses facial recognition technology (e.g., OpenCV's facial recognition module) to identify speakers from video data. The input is video data, and the output is data on the identified speaker. Specifically, it detects facial features in the video frame and compares them with an existing database to identify individuals.

[1020] Step 6:

[1021] The server uses emotion recognition technology (e.g., Affectiva SDK) to recognize participants' emotions from the collected audio and video data. The input is audio and video data, and the output is emotional state data. Specifically, it performs voice tone analysis and facial expression analysis to identify emotional levels.

[1022] Step 7:

[1023] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to automatically generate minutes in a specified format. The input is a dataset of each analysis result, and the output is the integrated minutes. Specifically, each piece of data is organized chronologically and a document is generated in a specified format.

[1024] Step 8:

[1025] The server stores the generated minutes in a database and provides them in a format that can be searched later. The input is the generated minutes, and the output is a searchable database. Specifically, the minutes data is indexed and stored in the database, making it searchable later using queries.

[1026] Step 9:

[1027] The server monitors the work environment in real time and detects the fatigue and stress levels of workers. The input is video, audio, and emotional data, and the output is data on the fatigue and stress levels of workers. Specifically, it analyzes continuously collected data and generates an alert if a certain threshold is exceeded.

[1028] In this way, it is possible to monitor participants' emotions and work environment in real time, along with detailed records of meetings and work content, and take appropriate measures.

[1029] 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.

[1030] 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.

[1031] 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.

[1032] [Fourth embodiment]

[1033] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1034] 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.

[1035] 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).

[1036] 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.

[1037] 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.

[1038] 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).

[1039] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1040] 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.

[1041] 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.

[1042] 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.

[1043] 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.

[1044] 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.

[1045] 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."

[1046] This invention is a system that analyzes the audio and video of a meeting and automatically generates detailed minutes. This system processes audio and video data, accurately capturing the content of materials and participants' remarks, and creates minutes.

[1047] System Program Overview

[1048] The system begins by having a user upload a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers. This data is then integrated to generate accurate minutes. The generated minutes are then stored on the server and made available in a searchable format.

[1049] Audio data processing

[1050] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[1051] Video data processing

[1052] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[1053] Identifying the speaker

[1054] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[1055] Generate meeting minutes

[1056] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in the specified format, which contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[1057] Providing generated minutes

[1058] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search for minutes based on "what a specific speaker said" or "a specific slide from the meeting."

[1059] Specific examples

[1060] For example, a user uploads a video of a "project progress meeting" to the server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify speakers, and this information is integrated to generate minutes. The generated minutes include details such as "Yamada spoke about important point A while explaining slide 3." These minutes are stored on the server and can be searched and viewed by users later.

[1061] This system allows detailed recording of meeting contents, allowing users to easily check and reuse the information later.

[1062] The processing flow will be explained below.

[1063] Step 1:

[1064] The user uploads the video file of the conference from the terminal to the server.

[1065] Step 2:

[1066] The server receives the uploaded video file and stores it in a temporary storage area.

[1067] Step 3:

[1068] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[1069] Step 4:

[1070] The server uses a speech recognition module (e.g., Google Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[1071] Step 5:

[1072] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[1073] Step 6:

[1074] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[1075] Step 7:

[1076] The server uses a face recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected face data is matched with the audio data.

[1077] Step 8:

[1078] The server combines the results of speech recognition, video analysis, and speaker identification to generate a draft of the minutes, which includes the text data of each statement, slide content, pointer movement, and speaker information.

[1079] Step 9:

[1080] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[1081] Step 10:

[1082] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[1083] Example 1

[1084] 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."

[1085] Accurately analyzing the audio and video of a meeting and automatically generating detailed minutes is a time-consuming and labor-intensive process. Furthermore, it is often difficult to accurately identify speakers and record slide content. Furthermore, it is difficult to quickly search through the generated minutes to find specific information.

[1086] 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.

[1087] In this invention, the server includes means for extracting audio data and converting it to text, means for dividing and analyzing video data by frame to detect slide content and pointer movement, means for identifying speakers using face recognition technology, means for generating minutes by integrating the results of the audio recognition, video analysis, and speaker identification, and means for saving the generated minutes and providing them in a searchable format, thereby enabling accurate recording of the meeting contents and rapid subsequent retrieval.

[1088] "Audio data" refers to digital data that is a recording of what is said during a meeting or the like.

[1089] "Video data" refers to digital data in the form of moving images captured during a conference.

[1090] A "voice recognition module" is a software or hardware component for converting voice data into text data.

[1091] A "frame" is a unit that refers to each still image that makes up video data.

[1092] A "video analysis module" is a software or hardware component that analyzes video data to detect slide content and pointer movement.

[1093] "Slide content" refers to the information on presentation slides displayed during a conference.

[1094] "Pointer movement" refers to the movement of the pointer or mouse on the slides during the meeting.

[1095] "Facial recognition technology" is a technology for identifying and recognizing individual faces from video data.

[1096] "Speaker" refers to a person who speaks during a meeting.

[1097] A "minutes" is a document that records in detail the progress of a meeting, what was said, the contents of slides, etc.

[1098] A "searchable format" is a data format that is indexed so that a user can later search for specific content.

[1099] The present invention is a system for automatically generating detailed meeting minutes by analyzing audio and video of a meeting. The system uses multiple hardware and software modules to process audio and video data and generate accurate meeting minutes.

[1100] Basic configuration

[1101] Audio data processing

[1102] The system begins operation when a user uploads a video file of a meeting to the server. The server first extracts audio data from the video file using a tool such as FFmpeg. This extracted audio data is then sent to a speech recognition module (e.g., Google Cloud Speech-to-Text API) and converted into text data. This text data is then used to record what was said by the meeting participants.

[1103] Video data processing

[1104] Next, the server splits the video file into frames using a tool such as FFmpeg. Each frame is then analyzed using a video analysis module such as OpenCV. This analysis detects the slide content and pointer movement.

[1105] Identifying the speaker

[1106] Facial recognition technology (e.g., Amazon Rekognition) is used to analyze the video data. The server uses the video data to detect the faces of each participant and identifies the speaker by matching them with the audio data. This makes it clear which participant made each comment.

[1107] Generate meeting minutes

[1108] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes in a specified format (e.g., Markdown or PDF). These minutes contain detailed records of the progress of the meeting, the contents of the materials, and the remarks of each speaker.

[1109] Providing generated minutes

[1110] The generated minutes are saved on the server. The saved minutes are indexed using Elasticsearch and provided in a format that makes them easy for users to search later. Users can access the minutes using a web browser or a dedicated app, and search and view them based on "the content of a specific speaker" or "a specific slide."

[1111] Specific examples

[1112] For example, a user uploads a video file of a "project progress meeting" to the server. This video file can be in supported formats such as MP4 or AVI. The server first extracts the audio data and converts it to text using the Google Cloud Speech-to-Text API. Next, the server divides the video data into frames and analyzes the slide content and pointer movement using OpenCV. It identifies speakers using facial recognition technology (e.g., Amazon Rekognition), and combines this data to generate minutes. The generated minutes include details such as "Participant A spoke about an important point while explaining slide 3." The minutes are stored on the server and can be searched and viewed later by users.

[1113] Prompt Sentence Examples

[1114] I have uploaded a video of a "Project Status Meeting." Please generate minutes of this meeting. The minutes should include the content of each slide and what each speaker said.

[1115] This system automatically and accurately records detailed meeting content, allowing users to quickly search for and check the information they need later.

[1116] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1117] Step 1:

[1118] The user uploads the video file of the conference to the server. At this stage, the video file is sent to the server and saved in the specified directory. The input is the video file provided from the user's device, and the output is the video file saved on the server side.

[1119] Step 2:

[1120] The server receives the video file and extracts the audio data. Using a tool such as FFmpeg, it separates the audio track from the video file. The input is the video file saved in step 1, and the output is an audio file temporarily stored on the server.

[1121] Step 3:

[1122] The server sends the audio data to the speech recognition module. The audio file is converted into text data using a method such as the Google Cloud Speech-to-Text API. During this process, the speech recognition module is called by the server. The input is the audio file, and the output is text data.

[1123] Step 4:

[1124] The server receives the text data from the speech recognition module and stores it as the speech content. This allows each speech made during the meeting to be recorded in text format. The input is the text data of the speech recognition results, and the output is the storage of the text data as the speech content.

[1125] Step 5:

[1126] The server splits the video file into frames. FFmpeg is used to convert the video into frames, which are a series of still images. The input is the video file saved in step 1, and the output is an image file split into frames.

[1127] Step 6:

[1128] The server sends the image files divided into frames to a video analysis module for analysis. Using a video analysis module such as OpenCV, slide content and pointer movement are detected from each frame. The input is the frame image file, and the output is the analysis results: slide content and pointer movement information.

[1129] Step 7:

[1130] The server uses facial recognition technology to identify the speaker. Using Amazon Rekognition or similar, it detects the face of each participant from the video data and compares it with the audio data. The input is the frame image file and audio data, and the output is data identifying the speaker.

[1131] Step 8:

[1132] The server combines the results of speech recognition, video analysis, and speaker identification to generate minutes. The minutes are created in the specified format (Markdown, PDF, etc.). The input is the data obtained at each processing step (text data, analysis results, speaker information), and the output is the completed minutes file.

[1133] Step 9:

[1134] The generated minutes are stored on a server and provided in a searchable format. They are indexed using a search engine such as Elasticsearch, allowing users to easily access and search them later. The input is the generated minutes file, and the output is an indexed database.

[1135] Step 10:

[1136] Users search and view meeting minutes using a web browser or a dedicated app. The server returns indexed meeting minutes in response to the user's request. The input is the user's search query, and the output is the relevant part of the meeting minutes as a search result.

[1137] (Application example 1)

[1138] 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."

[1139] Conventional meeting minutes systems require manual processing of audio and video data to create minutes, requiring a great deal of time and effort. Similar problems existed in online lectures and seminars, making it difficult to accurately record the content of lectures and details of Q&A sessions. Furthermore, searching this data later required a great deal of effort, preventing sufficient reuse of information.

[1140] 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.

[1141] In this invention, the server includes means for extracting audio data and converting it to text, means for analyzing video data to detect slide content and pointer movement, means for identifying speakers using facial recognition technology, means for integrating audio and video data to generate minutes, means for saving the generated minutes and making them searchable, means for extracting audio data from video, means for transmitting the extracted audio data to a voice recognition system and converting it to text data, and means for recording the content of lectures and question times. This allows the content of lectures and meetings to be automatically and accurately recorded, making it possible to easily search and reuse the information later.

[1142] "Audio data" refers to audio data recorded during a conference or lecture.

[1143] "Text data" is character string data converted from voice data using voice recognition technology.

[1144] "Video data" refers to video frame data recorded during a conference or lecture.

[1145] "Slide content" refers to information on presentation materials used during lectures and meetings.

[1146] "Pointer movement" refers to the movement trajectory of the cursor or pointer indicated on the screen during a presentation.

[1147] "Facial recognition technology" is a technology that identifies the face of a specific person from video data.

[1148] "Speaker" refers to the person who made a statement at a meeting or lecture.

[1149] "Integration" means combining individual audio data, video data, and speaker information into a single minutes.

[1150] "Minutes" are documents that record the progress of a meeting or lecture, what was said, slide information, etc.

[1151] "Searchable" means that stored information can be easily found based on specified criteria.

[1152] A "video" is a data file in which video and audio are recorded together.

[1153] A "voice recognition system" is a system that analyzes voice data and converts it into text data.

[1154] "Lecture content" refers to a series of statements, slides, and content intended for educational or explanatory purposes.

[1155] "Question windows" refer to specific time slots during lectures or meetings when questions arise.

[1156] This invention is a system that analyzes audio and video data from lectures and seminars and automatically generates detailed minutes. This system is realized using hardware such as a server, smartphone, and head-mounted display, and software such as a voice recognition module, video analysis module, and facial recognition technology.

[1157] System Overview

[1158] The system starts by a user uploading a video of a lecture or meeting to the server, which receives the video and performs the following steps:

[1159] 1. Extract audio data from video.

[1160] 2. The extracted voice data is sent to a voice recognition system and converted into text data.

[1161] 3. Analyze the video data to detect the slide content and pointer movement.

[1162] 4. Use facial recognition technology to identify speakers.

[1163] 5. Integrate these data to generate detailed meeting minutes.

[1164] 6. The generated minutes are saved on the server and made searchable.

[1165] Hardware and software used

[1166] Hardware:

[1167] Smartphones and head-mounted displays: Used to record and upload videos.

[1168] Server: Used to analyze data and generate transcripts.

[1169] software:

[1170] Python: Implementation of the main program and use of libraries.

[1171] cv2 (OpenCV): Video data analysis and frame segmentation.

[1172] speech_recognition: Converting speech data to text.

[1173] face_recognition: Application of facial recognition technology.

[1174] moviepy: Extracting audio data from videos.

[1175] transformers (Hugging Face): Speech recognition using the Wav2Vec2 model.

[1176] Processing flow

[1177] 1. Extracting audio data:

[1178] The server extracts the audio data from the uploaded video, which it does using the moviepy library.

[1179] 2. Speech Recognition:

[1180] The extracted audio data is converted into text data by a speech recognition system using the speech_recognition library. At this stage, the lecture content and speech content are obtained as text information.

[1181] 3. Video Data Analysis:

[1182] The server splits the video data into frames, detects the slide content and pointer movement using the OpenCV library, and identifies the speaker by recognizing the faces in each frame using the face_recognition library.

[1183] 4. Integration and Transcription:

[1184] The speech recognition results, video analysis results, and speaker information obtained at each step are integrated to generate minutes in the specified format. For example, the minutes will include information such as "Speaker A spoke about important point A while explaining slide 3."

[1185] 5. Archive and search meeting minutes:

[1186] The generated minutes are saved on the server for users to view and search later. Users can easily search for minutes based on "the content of a specific speaker" or "a specific slide."

[1187] Specific examples

[1188] For example, when a user uploads a video of an online lecture to the server, the system automatically analyzes the audio and video and generates detailed minutes that record the lecture content and the time periods when questions were asked. These minutes include detailed information such as, "Speaker A began explaining slide 1 immediately after the lecture began."

[1189] Prompt Sentence Examples

[1190] "Please convert the following audio data into text. Link to audio data: {URL of audio data}. Please describe the audio in chronological order and include speaker information if available."

[1191] As described above, the present invention makes it possible to automatically record the contents of lectures and meetings, and to easily search and reuse them later.

[1192] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1193] Step 1:

[1194] A user uploads video files of lectures or meetings to the server. The user selects the video files using their device and sends them to a specific upload destination on the server. The server stores the received video files and prepares them for further processing.

[1195] Input: Video file uploaded by user.

[1196] Output: Video file saved on the server.

[1197] Step 2:

[1198] The server extracts the audio data from the stored video file. For this process, it uses the moviepy library. The server reads the video file, separates the audio track, and saves it as a new audio file.

[1199] Input: Video file.

[1200] Output: Audio file.

[1201] Step 3:

[1202] The server sends the extracted audio file to a speech recognition module and converts it into text data. This process uses the speech_recognition library. The server reads the audio file and passes it through a speech recognition model to obtain text information.

[1203] Input: Audio file.

[1204] Output: Text data.

[1205] Step 4:

[1206] The server splits the video data from the video file into frames and analyzes them. This process uses the OpenCV library. The server splits the video file into frames and identifies the slide content, pointer movement, and speaker for each frame.

[1207] Input: Video file.

[1208] Output: Video data for each frame, analysis results (slide content, pointer movement, speaker information).

[1209] Step 5:

[1210] The server identifies the speaker based on the facial information acquired for each frame. Using the face_recognition library, the server compares the facial data in each frame with known facial data to identify the speaker.

[1211] Input: Face data for each frame.

[1212] Output: Identified speaker information.

[1213] Step 6:

[1214] The server then combines the acquired text data, slide content, pointer movement, and speaker information to generate detailed minutes. By combining each piece of data, the minutes are compiled in a format that users can easily review later.

[1215] Input: text data, slide content, pointer movement, speaker information.

[1216] Output: Detailed meeting transcript.

[1217] Step 7:

[1218] The server stores the generated minutes and provides them to users in a searchable format. The minutes are stored in a database and indexed so that users can easily search for specific criteria.

[1219] Input: Detailed minutes.

[1220] Output: Saved, searchable meeting minutes.

[1221] By using the above processing steps, the user can automatically record the contents of a lecture or conference, and later easily check the overall progress or the content of specific comments.

[1222] 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.

[1223] This invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates minutes of the meeting with an emotion engine that recognizes the emotions of users. This system processes audio and video data and utilizes the emotion engine to record the emotions of participants and reflect them in the minutes, making it possible to comprehensively understand the detailed situation of the meeting.

[1224] System Program Overview

[1225] The system begins by a user uploading a video of a meeting to the server. The server receives the video, extracts audio data and converts it into text, analyzes the video data to detect slide content and pointer movement, and uses facial recognition technology to identify speakers and an emotion engine to recognize participants' emotions. This data is then integrated to generate accurate minutes, which are then stored on the server and made available in a searchable format.

[1226] Audio data processing

[1227] When a user uploads a video, the server first extracts the audio data from the video, which is then sent to a speech recognition module and converted into text data, which is then used to record what the participants said.

[1228] Video data processing

[1229] The server then processes the video data, which is split into frames from the video. A video analysis module analyzes each frame to detect slide content and pointer movement. This information is used to record the details of the materials used in the meeting.

[1230] Identifying the speaker

[1231] Facial recognition technology is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker. This makes it clear which participant made each comment.

[1232] Emotion recognition

[1233] Using the emotion engine, a distinctive feature of the present invention, the server recognizes participants' emotions from the video and audio data. The emotion engine uses techniques such as facial expression analysis and audio tone analysis to identify the emotional state (e.g., joy, anger, sadness, etc.) of each speaker. The recognized emotion information is saved as text data.

[1234] Generate meeting minutes

[1235] The server combines the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes in a specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[1236] Providing generated minutes

[1237] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[1238] Specific examples

[1239] For example, a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text through a speech recognition module. Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Facial recognition technology is used to identify the speaker, and an emotion engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Yamada spoke about important point A while explaining slide 3, and at the same time, Yamada seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[1240] This system not only records the details of the meeting, but also grasps the emotional state of the participants, allowing for deeper understanding and analysis.

[1241] The processing flow will be explained below.

[1242] Step 1:

[1243] The user uploads the video file of the conference from the terminal to the server.

[1244] Step 2:

[1245] The server receives the uploaded video file and stores it in a temporary storage area.

[1246] Step 3:

[1247] The server uses libraries such as FFmpeg to extract audio data from the stored video files, and converts the extracted audio data into formats such as WAV.

[1248] Step 4:

[1249] The server uses a speech recognition module (e.g., Google Cloud Speech-to-Text API) to convert the voice data into text data, which is then temporarily stored.

[1250] Step 5:

[1251] The server divides the video file into frames at regular intervals and obtains the frame images, which are then sent to an image analysis module (e.g., OpenCV).

[1252] Step 6:

[1253] The image analysis module detects the slide content and pointer position from each frame image. The detected slide content and pointer movement are returned to the server as text data and coordinate data and temporarily saved.

[1254] Step 7:

[1255] The server uses a facial recognition module (e.g., FaceNet) to detect the faces of conference participants using the video data, and the detected facial data is matched with the audio data.

[1256] Step 8:

[1257] The server uses an emotion engine to recognize participants' emotions from video and audio data. The emotion engine analyzes facial expressions and audio tones, and stores the emotional state as text data.

[1258] Step 9:

[1259] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate a draft of the minutes, which includes text data for each statement, slide content, pointer movement, speaker information, and emotion information.

[1260] Step 10:

[1261] The server converts the generated draft minutes into a specified format (e.g., PDF or Word), which makes the minutes available for users to view and download.

[1262] Step 11:

[1263] The server stores the final minutes in a database and creates an index for future searches. Users can access the server's search interface through their browsers to search and view past minutes using specific keywords.

[1264] Example 2

[1265] 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."

[1266] Conventional meeting recording systems require a great deal of time and effort to process audio and video data separately and generate minutes manually. Furthermore, there is no way to grasp the emotional state of participants, making it difficult to understand the detailed situation of the meeting or the emotions of the participants. Furthermore, the generated minutes are not easily searchable or categorized, making them difficult to reference later.

[1267] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for extracting audio data and converting it into text, means for analyzing video data to detect visual information and indicator movements, means for identifying the caller using biometric recognition technology, means for recognizing an emotional state from the audio data and video data using a recognition engine, means for generating minutes by integrating the audio data, video data, caller information, and emotional information, and means for saving the generated minutes and making them categorizable. This enables audio and video to be processed in an integrated manner, automatically generating minutes that also include the caller's emotional state, and enabling efficient search and reference.

[1268] "Audio data" refers to recordings of sounds collected during meetings, conversations, etc.

[1269] "Text" is character information generated by processing audio data.

[1270] "Video data" refers to video information containing visual content such as meetings and presentations.

[1271] "Visual information" refers to information such as slide content and screen display analyzed from video data.

[1272] "Pointer movement" refers to interactive elements such as pointer or hand movements detected within the video data.

[1273] "Biometric recognition technology" is a technology that identifies individuals based on video data, such as facial recognition.

[1274] A "caller" is a person participating in a meeting or conversation.

[1275] A "cognitive engine" is a technology that analyzes audio and video data to recognize emotional states and other perceptual information.

[1276] "Emotional state" is information that indicates the psychological reactions and emotions (joy, anger, sadness, etc.) of the speaker or participants.

[1277] "Integration" is the process of combining various data to create a single, continuous record.

[1278] A "minutes" is a document that summarizes the contents of a meeting, what was said, the identity of the speakers, and their emotional state.

[1279] "Classifiable" means that the generated minutes are organized according to specific criteria or categories so that they can be easily searched and referenced.

[1280] A "system" is a structure having a series of integrated functions that includes all of the above means.

[1281] The present invention combines a system that analyzes the audio and video of a meeting, accurately grasps the content of materials and participants' remarks, and automatically generates meeting minutes with a recognition engine that recognizes user emotions. An embodiment of the present invention will be described in detail below.

[1282] The system starts by having a user upload a video of a meeting to the server, which receives the video and processes it in the following steps:

[1283] First, the server uses a library such as "FFmpeg" to extract audio data from the video. The extracted audio data is sent to a "speech recognition module" (general name) and converted into text data. For this purpose, a "speech recognition API" (Google Cloud Speech-to-Text API) or similar is used. This text data is then used to record what the participants are saying.

[1284] Next, the server divides the video into frames and processes the video data. A "video analysis module" (OpenCV) is used to divide the frames. The video analysis module analyzes each frame and detects the contents of the slides and the movement of the pointer. This information is used to record the details of the materials used in the meeting. In addition, "biometric recognition technology" (facial recognition technology) is used to analyze the video data. The server uses the video data to detect the faces of each participant and compares them with the audio data to identify the speaker.

[1285] An important feature is that the server uses a recognition engine, which includes technologies such as facial expression analysis and voice tone analysis (IBM Watson Tone Analyzer). Using these technologies, the server recognizes the emotional state (e.g., joy, anger, sadness, etc.) of each speaker from the video and voice data. The recognized emotional information is saved as text data.

[1286] Once this data is collected, the server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition, and generates minutes in the specified format. These minutes contain a detailed record of the progress of the meeting, the contents of the materials, and the remarks and emotions of each speaker.

[1287] The generated minutes are stored on a server and provided in a format that users can later access and search. For example, users can search minutes based on "the content of a specific speaker's speech and their emotions at the time" or "the emotional reaction to a specific slide in the meeting."

[1288] As a concrete example, imagine a scenario where a user uploads a video of a "project progress meeting" to a server. The server extracts audio data from the video and converts it into text using a "speech recognition API." Next, the video data is divided into frames and the slide content and pointer movement are analyzed. Biometric recognition technology is used to identify the speaker, and a recognition engine is used to recognize the speaker's emotions. The minutes generated by integrating this information include details such as "Person A spoke about important point A while explaining slide 3, and at the same time, Person A seemed nervous." These minutes are stored on the server and can be searched and viewed by users later.

[1289] An example prompt might be, "Upload a video of a project status meeting and automatically generate minutes. Please also record what Person A said and their emotions at the time."

[1290] The above is a specific description of the embodiment of the present invention. The present invention not only records the details of the meeting but also grasps the emotional states of the participants, making it possible to comprehensively understand and analyze the detailed situation of the meeting.

[1291] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1292] Step 1:

[1293] A user uploads a video of the meeting.

[1294] The user selects the video file of the "Project Progress Meeting" from their device and sends it to the server through the system's upload interface. As input, the video file of the meeting is passed from the device to the server. As output, the video file is saved on the server.

[1295] Step 2:

[1296] The server extracts the voice data and sends it to a voice recognition module.

[1297] The server analyzes the uploaded video file and extracts the audio data using a library such as FFmpeg. This audio data is then sent to a speech recognition API. The audio data extracted from the video file is used as input. The audio data is converted to text as output.

[1298] Step 3:

[1299] The server divides the video data into frames and processes them in a video analysis module.

[1300] The server splits the video data into frames using OpenCV. For each frame, it uses OCR technology to recognize the slide content and analyzes the pointer, hand, and other indicator movements. The video data is used as input. Data about the slide content and indicator movements is generated as output.

[1301] Step 4:

[1302] The server uses facial recognition technology to identify the speaker.

[1303] The server detects facial images in each frame of video data and identifies the speaker using facial recognition technology such as Amazon Rekognition. The detected facial images are then matched with the audio data. The facial images and audio data are used as input. The output is data that identifies the speaker.

[1304] Step 5:

[1305] The server uses an emotion engine to recognize the emotions of the participants.

[1306] The server uses an emotion engine to analyze the video and audio data and recognize the speaker's emotional state. This process uses facial expression analysis and tone analysis techniques. The video and audio data are used as input. The output is an emotion recognition result (e.g., joy, anger, sadness).

[1307] Step 6:

[1308] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to generate minutes.

[1309] The server integrates the data collected so far and generates minutes in the specified format. These minutes include detailed records of the progress of the meeting, the contents of the materials, and the remarks and emotions of the speakers. Various analysis and recognition results are used as input. The integrated minutes data is generated as output.

[1310] Step 7:

[1311] The server provides the generated minutes.

[1312] The generated minutes are stored on a server and provided in a format that users can access later for searching and viewing. Specifically, users can use the search interface to search for "emotional reactions to a specific slide in a meeting," "the content of a specific speaker's remarks and their emotions at the time," and so on. The user's search query is used as input, and the relevant minutes data is provided as output.

[1313] The above is the specific processing flow of the system program divided into specific processing steps, and the details of each process accompanied by input and output.

[1314] (Application example 2)

[1315] 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."

[1316] Conventional meeting recording systems convert audio data into text and analyze video data, but lack the functionality to monitor participants' emotions and work environment in real time. This makes it difficult to understand the details of the meeting content or to properly grasp the fatigue and stress levels of workers. Furthermore, there is a need for a system that can respond in real time to changes in the work environment.

[1317] 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 extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movement, means for identifying the speaker using face recognition technology, means for recognizing the emotions of participants using emotion recognition technology, means for integrating audio data and video data to generate minutes, means for saving the generated minutes and making them searchable, and means for monitoring the work environment in real time and detecting the fatigue and stress of workers. This makes it possible to not only record detailed information about meetings and work content, but also to monitor the emotions of participants and the work environment in real time.

[1318] "Voice data" refers to voice information such as statements and instructions that occur during meetings or work.

[1319] "Means for converting to text" refers to the technology or equipment that converts voice data into text information.

[1320] "Video data" refers to video information captured by cameras or recording devices.

[1321] "Slide content" refers to the information on slides used in presentations or as meeting materials.

[1322] "Indicator movement" refers to the movement of a pointer, mouse cursor, or other device on the screen to indicate instructions or a point of interest.

[1323] "Facial recognition technology" refers to technology for recognizing and identifying specific individuals from video data.

[1324] "Means for identifying the speaker" refers to technologies and methods for determining who is speaking based on audio and video data.

[1325] "Emotion recognition technology" refers to technology for identifying an individual's emotional state (joy, anger, sadness, etc.) from audio and video data.

[1326] "Means for recognizing emotions of participants" refers to a technique or method for determining the emotions of participants in a meeting or workplace using emotion recognition technology.

[1327] "Means for generating minutes" refers to technologies and methods that integrate audio data, video data, emotional information, etc. to automatically generate documents that record the content of meetings and work.

[1328] "Means for storing and making searchable minutes" refers to the technology and methods for electronically storing the generated minutes in a database so that they can be searched and viewed at a later date.

[1329] "Means for monitoring the work environment in real time" refers to technologies and methods for instantly monitoring the progress and environment of work in factories and workplaces.

[1330] "Means for detecting worker fatigue and stress" refers to technologies and methods for assessing the health and mental state of workers using emotion recognition technology or other sensor technology.

[1331] The present invention relates to a work monitoring and quality control system using factory robots. The system includes means for extracting audio data and converting it into text, means for analyzing video data to detect slide content and indicator movements, means for identifying speakers using facial recognition technology, means for recognizing participants' emotions using emotion recognition technology, means for integrating audio data and video data to generate minutes of meetings, means for storing the generated minutes and making them searchable, and means for monitoring the work environment in real time to detect worker fatigue and stress.

[1332] System Program Overview

[1333] The server collects video and audio data from within the factory using high-resolution cameras and microphones. The collected audio data is converted into text using a speech recognition module (e.g., IBM Watson Speech to Text). The video data is divided into frames and analyzed using OpenCV. The speaker is identified using slide content, indicator movements, and facial recognition technology.

[1334] Using emotion recognition technology (e.g., Affectiva SDK), the emotions of workers are recognized from video and audio data. This allows participants' emotional information to be saved as text data. The results of audio recognition, video analysis, speaker identification, and emotion recognition are integrated to generate minutes in the specified format.

[1335] Hardware and Software Use

[1336] The system uses a high-resolution camera, microphone, high-performance CPU and GPU, and utilizes TensorFlow (machine learning library), OpenCV (image processing library), IBM Watson Speech to Text (voice recognition), and Affectiva SDK (emotion recognition) software.

[1337] Specific examples

[1338] For example, if a factory worker says, "I can't assemble this part properly," the voice recording is collected and converted into text. Emotion recognition technology recognizes that the worker is feeling "frustrated." This information is collected and analyzed as follows:

[1339] Audio data:

[1340] "This part doesn't assemble properly" -> recognized_text: "This part doesn't assemble properly"

[1341] Emotional Data:

[1342] recognized_emotion: "frustration"

[1343] We also provide an example of input to a generative AI model using the following prompt sentence:

[1344] "Generate a monitoring report for the factory. Analyze the voice and emotions of the workers and incorporate them into the report."

[1345] This allows the system to record detailed factory working conditions as well as monitor workers' emotions and health in real time, allowing appropriate responses to be taken.

[1346] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1347] Step 1:

[1348] The server uses microphones to collect voice data from within the factory. This allows it to obtain voice information such as conversations and instructions within the factory. Specifically, the analog voice signals collected by the microphones are converted into digital data and input into a voice recognition module. The input is voice data, and the output is digitized voice data.

[1349] Step 2:

[1350] The server uses a speech recognition module (e.g., IBM Watson Speech to Text) to convert the collected voice data into text. The input is digital voice data, and the output is text data. Specifically, the voice signal is converted into acoustic features, which are then input into a model to generate a corresponding string of characters.

[1351] Step 3:

[1352] The server uses a high-resolution camera to collect video data within the factory, visually recording the progress and environment of work. The input is visual information and the output is video data. The video data collected by the camera is divided into frames and input into the image processing module.

[1353] Step 4:

[1354] The server divides the video data into frames and analyzes them using OpenCV. The input is the video data, and the output is analyzed data on the slide content and indicator movement. Specifically, image processing is performed on each frame to extract the characteristics of the slides and indicators.

[1355] Step 5:

[1356] The server uses facial recognition technology (e.g., OpenCV's facial recognition module) to identify speakers from video data. The input is video data, and the output is data on the identified speaker. Specifically, it detects facial features in the video frame and compares them with an existing database to identify individuals.

[1357] Step 6:

[1358] The server uses emotion recognition technology (e.g., Affectiva SDK) to recognize participants' emotions from the collected audio and video data. The input is audio and video data, and the output is emotional state data. Specifically, it performs voice tone analysis and facial expression analysis to identify emotional levels.

[1359] Step 7:

[1360] The server integrates the results of speech recognition, video analysis, speaker identification, and emotion recognition to automatically generate minutes in a specified format. The input is a dataset of each analysis result, and the output is the integrated minutes. Specifically, each piece of data is organized chronologically and a document is generated in a specified format.

[1361] Step 8:

[1362] The server stores the generated minutes in a database and provides them in a format that can be searched later. The input is the generated minutes, and the output is a searchable database. Specifically, the minutes data is indexed and stored in the database, making it searchable later using queries.

[1363] Step 9:

[1364] The server monitors the work environment in real time and detects the fatigue and stress levels of workers. The input is video, audio, and emotional data, and the output is data on the fatigue and stress levels of workers. Specifically, it analyzes continuously collected data and generates an alert if a certain threshold is exceeded.

[1365] In this way, it is possible to monitor participants' emotions and work environment in real time, along with detailed records of meetings and work content, and take appropriate measures.

[1366] 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.

[1367] 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.

[1368] 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.

[1369] 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.

[1370] 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.

[1371] 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.

[1372] 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).

[1373] 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.

[1374] 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."

[1375] 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.

[1376] 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).

[1377] 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.

[1378] 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.

[1379] 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.

[1380] 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.

[1381] 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.

[1382] 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.

[1383] 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.

[1384] 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.

[1385] 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.

[1386] 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.

[1387] The following is further disclosed regarding the above embodiment.

[1388] (Claim 1)

[1389] a means for extracting and converting the audio data into text;

[1390] means for analyzing the video data to detect the slide contents and the movement of the pointer;

[1391] A means for identifying a speaker using facial recognition technology;

[1392] a means for integrating audio data and video data to generate minutes;

[1393] a means for storing and making searchable the generated minutes;

[1394] A system including:

[1395] (Claim 2)

[1396] 10. The system of claim 1, which uses a voice recognition module to extract the voice data.

[1397] (Claim 3)

[1398] 10. The system of claim 1, further comprising means for dividing the video data into frames for analysis.

[1399] "Example 1"

[1400] (Claim 1)

[1401] a means for extracting and converting the audio data into text;

[1402] means for dividing and analyzing the video data into frames to detect the slide contents and pointer movement;

[1403] A means for identifying a speaker using facial recognition technology;

[1404] a means for generating minutes by integrating the results of speech recognition, video analysis, and speaker identification;

[1405] a means for storing the generated minutes and providing them in a searchable format;

[1406] A system including:

[1407] (Claim 2)

[1408] 10. The system of claim 1, wherein the system uses a voice recognition module to extract the voice data.

[1409] (Claim 3)

[1410] 10. The system of claim 1, further comprising means for dividing the video data into frames for analysis.

[1411] "Application Example 1"

[1412] (Claim 1)

[1413] a means for extracting and converting the audio data into text;

[1414] means for analyzing the video data to detect the slide contents and the movement of the pointer;

[1415] A means for identifying a speaker using facial recognition technology;

[1416] a means for integrating audio data and video data to generate minutes;

[1417] a means for storing and making searchable the generated minutes;

[1418] means for extracting audio data from the video;

[1419] means for transmitting the extracted voice data to a voice recognition system and converting the voice data into text data;

[1420] A means of recording lecture content and question times,

[1421] A system including:

[1422] (Claim 2)

[1423] 10. The system of claim 1, which uses a voice recognition module to extract the voice data.

[1424] (Claim 3)

[1425] 10. The system of claim 1, further comprising means for dividing the video data into frames for analysis.

[1426] "Example 2: Combining Emotion Engines"

[1427] (Claim 1)

[1428] a means for extracting and converting the audio data into text;

[1429] means for analyzing the video data to detect visual information and indicator movement;

[1430] A means for identifying the caller using biometric recognition technology;

[1431] means for recognizing an emotional state from audio data and video data using a cognitive engine;

[1432] a means for generating minutes by integrating audio data, video data, sender information, and emotion information;

[1433] A means for storing and categorizing the generated minutes;

[1434] A system including:

[1435] (Claim 2)

[1436] 10. The system of claim 1, which uses a voice recognition module to extract the voice data.

[1437] (Claim 3)

[1438] 10. The system of claim 1, further comprising means for dividing the video data into frames for analysis.

[1439] "Application example 2 when combining emotion engines"

[1440] (Claim 1)

[1441] a means for extracting and converting the audio data into text;

[1442] means for analyzing the video data to detect the slide content and the indicator movement;

[1443] A means for identifying a speaker using facial recognition technology;

[1444] a means for recognizing emotions of participants using emotion recognition technology;

[1445] a means for integrating audio data and video data to generate minutes;

[1446] a means for storing and making searchable the generated minutes;

[1447] A means of monitoring the work environment in real time and detecting worker fatigue and stress;

[1448] A system including:

[1449] (Claim 2)

[1450] 10. The system of claim 1, which uses a voice recognition module to extract the voice data.

[1451] (Claim 3)

[1452] 10. The system of claim 1, further comprising means for dividing the video data into frames for analysis. [Explanation of symbols]

[1453] 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. a means for extracting and converting the audio data into text; means for analyzing the video data to detect the slide contents and the movement of the pointer; A means for identifying a speaker using facial recognition technology; a means for integrating audio data and video data to generate minutes; a means for storing and making searchable the generated minutes; A system including:

2. 10. The system of claim 1, which uses a voice recognition module to extract the voice data.

3. 2. The system of claim 1, further comprising means for dividing the video data into frames for analysis.

Images