System

Abstract

A system is provided.SOLUTION: A system comprising: means for data collection; means for preprocessing; means for training a generative AI model; means for user input; means for video generation; and means for user viewing.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] Because there was no general awareness of preserving television programs up until the early 1970s, much of the video and audio data is missing. As a result, programs that are valuable historical materials have been scattered and lost, and viewers today and in the future will lose the opportunity to watch them. There is a need for an effective means of archiving past television programs and passing them on to future generations. [Means for solving the problem]

[0005] To solve this problem, the present invention provides a system that includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing. The preprocessing means has the function of analyzing and filtering data to remove noise and unnecessary parts. Furthermore, the generative AI model has the function of converting audio data into text and matching it with a script to fill in missing parts. This series of processes creates a system that can recreate past programs and provide them in a form that can be viewed by current and future viewers.

[0006] "Means for collecting data" refers to components that have the function of searching for and acquiring the necessary data such as video, audio, scripts, etc. from the Internet, libraries, archive facilities, etc.

[0007] The "means for performing preprocessing" is a component that has the function of analyzing collected data, removing noise and unnecessary parts, and identifying missing parts.

[0008] A "means for training a generative AI model" is a component that uses collected and preprocessed data to build a generative AI model and learns to fill in any gaps.

[0009] The "means for user input" is a component that provides an interface for the user to input information about the program that the user wants to reproduce, and has the function of transmitting that information to the system.

[0010] The "means for generating an image" is a component that has the function of creating an image that fills in missing parts using information provided by the user and a generative AI model.

[0011] The "means for user viewing" is a component that provides streaming playback and download functions to enable users to view the generated video.

[0012] "Data analysis and filtering" refers to the process of removing noise and unnecessary parts in order to improve the quality of collected data.

[0013] "Converting voice data to text" refers to the process of converting voice information into text information, which allows matching with a script.

[0014] "Completing missing parts" refers to the process of using a generative AI model to reproduce missing parts of collected data, maintaining overall continuity. [Brief explanation of the drawings]

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

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

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

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

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

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

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

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

[0023] [First embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0036] This invention is a system for recreating and archiving past television programs using generative AI. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing.

[0037] Specific program processing

[0038] 1. Data Collection

[0039] The server collects the necessary data such as video, audio, scripts, etc. from the Internet, libraries, and archive facilities.

[0040] The server obtains data using APIs or scraping techniques and temporarily stores it.

[0041] 2. Data Preprocessing

[0042] The server analyzes the collected data and removes noise and unnecessary parts.

[0043] The server converts the audio data into text and matches it with the script to identify missing parts.

[0044] 3. Generative Model Training

[0045] The server uses the collected data to train a generative AI model, specifically using frameworks like TensorFlow and PyTorch.

[0046] The server evaluates the model's performance every epoch during the training process and makes readjustments as needed.

[0047] 4. User Input

[0048] The user inputs information about the program they want to reproduce (program name, broadcast year, cast, etc.) through the terminal interface.

[0049] The terminal converts the user's input into JSON format and sends it to the server.

[0050] 5. Image Generation

[0051] Based on a request from a user, the server searches the database for the corresponding program data.

[0052] The server uses a generative AI model to generate a video that fills in the missing parts.

[0053] The server encodes the generated video, converts it into a viewable format, and stores it in a database.

[0054] 6. User Viewing

[0055] The user watches the generated video through a device that offers streaming playback and download options.

[0056] When a user watches a video, the device uses a cache function to provide efficient access.

[0057] Specific examples

[0058] For example, imagine a user wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. The server then collects the relevant data and fills in the missing parts with a generative AI model. The resulting video can then be streamed or downloaded and viewed by the user. This process brings a precious past show back to life in a new form, making it available for both current and future audiences to enjoy.

[0059] In this way, the present invention realizes a system that can recreate past television programs and provide them in a viewable format. This system fully utilizes the capabilities of generative AI to fill in missing parts, maintaining the continuity and quality of the video, allowing viewers to access valuable video assets from the past.

[0060] The processing flow will be explained below.

[0061] Step 1:

[0062] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to automate the retrieval of information and efficiently extract the relevant content.

[0063] Step 2:

[0064] The server analyzes and pre-processes the collected data, specifically filtering to remove noise from the video data and clean up the audio data, and uses speech recognition technology to convert the audio data into text.

[0065] Step 3:

[0066] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[0067] Step 4:

[0068] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[0069] Step 5:

[0070] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[0071] Step 6:

[0072] The user uses the interface on their terminal to input information about the program they want to reproduce (title, broadcast year, cast, etc.), which sends a specific request tailored to the user's needs to the system.

[0073] Step 7:

[0074] The device converts the request entered by the user into JSON format and sends it to the server. The server analyzes the received request and searches the database for the corresponding program data.

[0075] Step 8:

[0076] The server uses a generative AI model to fill in the gaps and generate the complete video data, using technology to generate consecutive scenes in a consistent manner while maintaining synchronization between audio and video.

[0077] Step 9:

[0078] The server encodes the generated video into a viewable format (e.g., MP4) and stores it in a database, where it can be instantly streamed or downloaded.

[0079] Step 10:

[0080] Users view the generated video through a device that offers streaming playback and download options, allowing users to freely access the video.

[0081] Step 11:

[0082] When a user watches a video, the device uses a cache function to shorten the video loading time and provide smooth playback. Any user feedback is sent to the server to help improve the system.

[0083] In this way, the server, terminal, and user play their respective roles at each step, realizing a system for reproducing past television programs.

[0084] Example 1

[0085] 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."

[0086] Many past television programs have been lost over the years, resulting in the loss of valuable cultural assets for present and future viewers. Restoring lost programs using conventional technology has been difficult and often requires advanced expertise and equipment. As a result, there is a problem in that opportunities for the general public to easily watch these programs are limited.

[0087] 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.

[0088] In this invention, the server includes a means for collecting data, a means for analyzing and filtering the data, and a means for training a generative AI model, which makes it possible to complement missing parts of past TV programs and reproduce them in a viewable form.

[0089] "Means of collecting data" refers to the means of collecting the necessary data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[0090] "Means for analyzing and filtering data" refers to means for removing noise and unnecessary parts from collected data, converting voice data into text, and matching it with a script.

[0091] "Means for training a generative AI model" means means for utilizing collected and preprocessed data to train a generative AI model and obtain highly accurate output.

[0092] "Means for users to input information" refers to an interface for users to input information about the program they wish to reproduce (e.g., program name, broadcast year, cast, etc.) and means for processing the input content.

[0093] "Means for generating images" refers to means for generating images that fill in missing parts using a pre-processed and trained generative AI model.

[0094] "Means for storing the generated video and converting it into a viewable format" refers to means for encoding the video created by the generative AI model, converting it into a common viewing format, and storing it.

[0095] The "means for user viewing" refers to an interface and system that allows a user to view the generated video by streaming playback or downloading it.

[0096] This invention is a system for recreating and archiving past television programs using generative AI. This system includes means for collecting data, means for preprocessing, means for training a generative AI model, means for users to input information, means for generating video, means for saving the generated video and converting it into a viewable format, and means for users to view it.

[0097] The server first uses a data collection method to collect the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archive facilities. Specifically, it obtains the data using APIs (e.g., YouTube Data API or open databases) or scraping technologies (e.g., Scrapy or Selenium), and temporarily stores them.

[0098] The server then uses data analysis and filtering techniques to remove noise and unwanted parts from the collected data. At this stage, FFmpeg and OpenCV are used to process the video data, and noise reduction techniques are used to process the audio data. The audio data is also converted to text using the Google Cloud Speech-to-Text API, which is then matched against a script to identify missing parts.

[0099] The server then uses a method to train the generative AI model using the preprocessed dataset. Specifically, it uses TensorFlow or PyTorch to train the model and obtain highly accurate output. During the training process, the model's performance is evaluated after each epoch, and techniques such as early stopping are applied to prevent overfitting.

[0100] The user inputs the information of the program they want to reproduce through the terminal interface using the information input means. The input content is converted to JSON format and sent to the server using the HTTPS protocol. For example, the user inputs the following prompt sentence:

[0101] "Please recreate a segment from a variety show that aired in 1975. The cast is a well-known celebrity."

[0102] The server uses a video generation method to generate a continuous video based on a user request, filling in missing parts with a generative AI model. The generated video is then encoded and converted into a viewable MP4 format. Codecs used include H.264 and HEVC. The generated video is then stored in the server's database.

[0103] Finally, the user uses a viewing means to view the generated video on a device. The device provides streaming playback or download options and utilizes a streaming server to efficiently deliver the generated video. It also uses a streaming cache function while the user is viewing to provide a smooth viewing experience.

[0104] In this way, the present invention realizes a system that can reproduce past television programs and provide them in a viewable format, thereby enabling viewers to easily access valuable video assets from the past.

[0105] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0106] Program processing steps

[0107] Step 1: Data collection

[0108] Description: The server collects the necessary data (video, audio, scripts, etc.) from the internet, libraries, and archives, using APIs (e.g., open databases) and scraping techniques (e.g., Scrapy and Selenium).

[0109] Input: Raw data from the internet, libraries and archival facilities.

[0110] Data processing / data calculation: Data acquisition via API or scraping.

[0111] Output: Captured video data, audio data, and script data.

[0112] Step 2: Data Preprocessing

[0113] Description: The server analyzes the collected data and removes noise and unnecessary parts. Specifically, it uses FFmpeg and OpenCV for video data and applies noise reduction technology to audio data. It also converts audio data to text using the Google Cloud Speech-to-Text API and matches it with the script to identify missing parts.

[0114] Input: Captured video data, audio data, and script data.

[0115] Data processing / data calculation: noise removal, speech-to-text conversion, text-to-script matching.

[0116] Output: Preprocessed video data, text data, and missing part information.

[0117] Step 3: Generative model training

[0118] Description: The server trains a generative AI model using preprocessed data. Specifically, it uses TensorFlow or PyTorch to split the dataset into training data and validation data and train the model. It evaluates the model's performance every epoch and applies techniques such as early stopping to prevent overfitting.

[0119] Input: Preprocessed video data, text data, and missing part information.

[0120] Data processing / data computation: model training, performance evaluation, and hyperparameter tuning.

[0121] Output: A trained generative AI model.

[0122] Step 4: User Input

[0123] Description: The user inputs the information of the program they want to reproduce through the device interface. The device converts the user's input into JSON format and sends it to the server using the HTTPS protocol.

[0124] Input: Program information entered by the user (e.g., program name, broadcast year, cast).

[0125] Data processing / data calculation: Convert user input into JSON format and send to the server.

[0126] Output: The request data sent to the server.

[0127] Step 5: Image generation

[0128] Description: The server searches the database for the relevant program data based on the user's request, and generates a video that fills in the missing parts using a generative AI model.The generated video is then encoded using a codec such as H.264 or HEVC, converted to MP4 format, and saved in the database.

[0129] Input: User request data, trained generative AI model, and program data from the database.

[0130] Data processing / data calculation: database search, generation of complementary video, video encoding, conversion to MP4 format.

[0131] Output: The generated video data.

[0132] Step 6: User Viewing

[0133] Description: The user uses a device to watch the generated video. The device uses a streaming server (e.g., NGINX or Wowza) to stream the video or provides a download option. During viewing, a cache function is used to provide a smooth viewing experience.

[0134] Input: Generated video data.

[0135] Data processing / data calculation: streaming distribution, cache processing.

[0136] Output: The video that the user can see.

[0137] (Application example 1)

[0138] 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."

[0139] There is a need to recreate valuable past television programs and video assets so that they can be viewed by current and future viewers. However, over the years, the original video data can deteriorate or become missing. Furthermore, when recreating past video, noise and unnecessary information are often mixed in, significantly reducing viewing quality. Therefore, a system is needed that can efficiently recreate past television programs and provide them to viewers with high quality.

[0140] 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.

[0141] In this invention, the server includes means for collecting data, means for preprocessing, means for training a generative AI model, means for user input, means for generating video, means for user viewing, means for analyzing the collected data and processing it to reproduce video of a specified program, means for providing an interface for users to request past video, and means for converting video generated using hardware into a viewable format, thereby making it possible to reproduce past TV programs in high quality and efficiently provide video requested by viewers.

[0142] "Data collection means" refers to the devices and methods used to collect the necessary data from the Internet and archive facilities.

[0143] "Preprocessing means" refers to a device or method that analyzes collected data and removes noise and unnecessary parts.

[0144] A "means for training a generative AI model" is a device or method that uses collected data to train an AI model and acquire knowledge to fill in any gaps.

[0145] "User input means" refers to a device or method for a user to input information about a program that the user wants to reproduce (such as the program name, broadcast year, and cast members).

[0146] "Image generation means" refers to a device or method that uses preprocessed data and a generative AI model to generate images that fill in missing parts.

[0147] "User viewing means" refers to a device or method that allows a user to view the generated video.

[0148] The "collected data analysis means" refers to a device or method that analyzes the collected data and performs the processing necessary to reproduce the video of the specified program.

[0149] The "user interface providing means" refers to a device or method that provides an interface for a user to request past video.

[0150] "Video conversion means" refers to a device or method that converts video generated using hardware into a viewable format.

[0151] This invention is a system that uses generative AI to recreate and archive past television programs. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing. Specific embodiments of these means are described below.

[0152] First, in the data collection process, the server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. This is achieved using APIs and scraping technology. The collected data is temporarily stored and used for subsequent processing.

[0153] Next, the pre-processing means analyzes the data collected by the server and removes noise and unnecessary parts. Specifically, it performs frame analysis of the collected video data and converts audio data into text. This converts the data into a format suitable for video complementation.

[0154] The generative AI model training method uses the collected data to train the generative AI model. Using frameworks like TensorFlow or PyTorch, the model's performance is evaluated every epoch and retuned as needed. As the model learns, it acquires knowledge to fill in missing parts of the video.

[0155] The user input means allows the user to input information about the program they want to reproduce (such as the program name, broadcast year, and cast members) through the device's interface. The device converts the user's input into JSON format and sends it to the server, allowing the server to efficiently search for specific program data.

[0156] In the video generation method, the server uses the generative AI model to generate videos that fill in missing parts. These videos are encoded into a format that can be viewed by users and stored in a database.

[0157] Finally, as a means for user viewing, users can view the generated video through their devices, with streaming playback and download options available, and a cache function for efficient access.

[0158] As a concrete example, consider a user who wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. In this case, the prompt sentence is "Please recreate the 1970 variety show 'Retro Variety Show'." The server collects the relevant data and fills in the missing parts with a generative AI model. As a result, the user can stream or download the generated video and watch it.

[0159] As described above, the present invention makes it possible to reproduce past television programs with high quality and efficiently provide the video requested by the viewer.

[0160] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0161] Step 1: Data collection

[0162] The server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. Specifically, it uses APIs and scraping technology to obtain data related to the specified program. The input is the program name and broadcast year, and the output is temporarily stored video, audio, and text data.

[0163] Step 2: Data Preprocessing

[0164] The server analyzes the collected data and processes it to remove noise and unnecessary parts. Specifically, it uses OpenCV to analyze frames of video data and applies a noise removal filter. Additionally, audio data is converted into text using speech recognition technology and matched with a script. The input is the raw collected data, and the output is the clean data that has been analyzed and filtered.

[0165] Step 3: Training the generative AI model

[0166] The server uses the preprocessed data to train a generative AI model. This uses frameworks such as TensorFlow and PyTorch. Specifically, it pairs video and text data to create a training set, and then trains the model based on that. The input is the preprocessed data, and the output is a fully trained generative AI model.

[0167] Step 4: User Input

[0168] The user inputs information about the program they want to play (such as the program name, broadcast year, and cast) through the device's interface. The device converts this input into JSON format and sends it to the server. The input is the user's program request information, and the output is JSON format data.

[0169] Step 5: Image generation

[0170] Based on a user request, the server uses a generative AI model to generate a video with missing parts filled in. Specifically, it uses a trained model to generate reconstructed video frames from clean data. The input is the user request and clean data, and the output is the filled video data.

[0171] Step 6: User Viewing

[0172] The user watches the generated video through a device. The device provides a player for streaming the video, and also offers a download option. The input is the generated video data, and the output is a viewable video.

[0173] As described above, by performing appropriate data processing and data calculations based on the input data at each step, a system is realized that reproduces past television programs with high quality and provides the video requested by the user.

[0174] 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.

[0175] This invention is a system that uses generative AI to recreate past television programs and combines it with an emotion engine that recognizes user emotions. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for user viewing, and an emotion engine that recognizes user emotions.

[0176] Specific program processing

[0177] 1. Data Collection

[0178] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives using APIs and scraping technology.

[0179] 2. Data Preprocessing

[0180] The server analyzes the collected data, filtering out noise and unnecessary parts, converts the audio data to text, and matches it with a script to identify missing parts, thereby preparing the data for subsequent input into a generative AI model.

[0181] 3. Generative Model Training

[0182] The server trains a generative AI model using the preprocessed data. Using the prepared dataset, it builds a generative AI model that fills in the gaps. The training uses the TensorFlow and PyTorch frameworks, and also adjusts hyperparameters.

[0183] 4. User Input

[0184] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The device converts the input information into JSON format and sends it to the server.

[0185] 5. Image Generation

[0186] The server receives the user's request, searches the database for the corresponding program data, uses a generative AI model to fill in the gaps and create a continuous video, and then encodes the video and converts it into a viewable format (e.g., MP4).

[0187] 6. Operation of the Emotion Engine

[0188] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. This information is taken into account when generating the video, and the parameters of the generative AI model are adjusted as needed.

[0189] 7. User Viewing

[0190] Users watch the generated video through the device. An emotion engine analyzes users' reactions in real time, and the content and tone of the video may be dynamically changed. The device offers streaming and download options to enhance the viewing experience.

[0191] Specific examples

[0192] For example, consider a user requesting a re-enactment of a 1970s variety show. If the emotion engine determines that the user is in a positive emotional state (e.g., joy), the generative AI model applies a filter that emphasizes more joyful scenes and brighter tones. On the other hand, if the user is in a sad emotional state, the video content and musical tone are adjusted to match that emotion. In this way, a customized video experience is provided based on the user's emotions.

[0193] This system not only makes precious programs from the past available to current and future audiences in new formats, but also customizes the visual experience based on the user's emotional state. The collaboration between generative AI and the emotion engine makes it possible to provide even more compelling visual content.

[0194] The processing flow will be explained below.

[0195] Step 1:

[0196] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to efficiently extract the relevant content.

[0197] Step 2:

[0198] The server analyzes the collected data and removes noise and unnecessary parts, specifically by performing noise reduction processing on video data and filtering on audio data, and also by using voice recognition technology to convert audio data into text.

[0199] Step 3:

[0200] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[0201] Step 4:

[0202] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[0203] Step 5:

[0204] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[0205] Step 6:

[0206] The user uses the interface on their device to input information about the program they want to replay (title, broadcast year, cast, etc.), and the system processes the request based on this information.

[0207] Step 7:

[0208] The device converts the request entered by the user into JSON format and sends it to the server, which then analyzes the request and searches the database for the corresponding program data.

[0209] Step 8:

[0210] The server uses a generative AI model to fill in the gaps and generate the complete video data, generating consecutive scenes in a consistent manner while maintaining audio and video synchronization.

[0211] Step 9:

[0212] The server encodes the generated video into a viewable format (e.g., MP4 format) and stores it in a database. The stored video can then be used for streaming or download.

[0213] Step 10:

[0214] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state, and the analysis results are sent to the server when the video is generated.

[0215] Step 11:

[0216] The server receives data from the emotion engine and adjusts the parameters of the generative AI model based on the user's emotional state, generating customized video according to the user's emotions.

[0217] Step 12:

[0218] The user watches the generated video through a device that provides streaming and download options to allow the user to watch the video in the appropriate way.

[0219] Step 13:

[0220] As users watch videos, the emotion engine analyzes their reactions in real time and transmits the information to the server, which can then dynamically adjust the content and tone of the video to provide the best possible viewing experience based on the user's emotions.

[0221] Example 2

[0222] 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."

[0223] In systems that reproduce past TV programs, there is a need for a method that not only plays back existing data but also provides a customized experience based on the user's emotions. Another challenge is to efficiently perform preprocessing such as noise removal from collected data and text conversion of audio data to achieve high-precision video generation.

[0224] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, and a means for user viewing. This makes it possible to reproduce past TV programs with high accuracy and dynamically change the video content and tone according to the user's emotions.

[0225] "Means of collecting data" refers to the means of obtaining data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[0226] "Preprocessing means" refers to the means for analyzing collected data, removing noise, deleting unnecessary parts, and preparing the data in a format suitable for a generative AI model.

[0227] "Means for training a generative AI model" refers to a means for training an AI model using preprocessed data and building a highly accurate model that can fill in missing parts.

[0228] The "means for user input" refers to a means by which a user inputs information about a program that the user wants to reproduce through a terminal and transmits the information to the server.

[0229] "Means for generating video" refers to a means for using a generative AI model to create video of a program requested by a user and to fill in missing parts to generate continuous video.

[0230] The "means for recognizing the user's emotions" is a means for analyzing the user's facial expressions, tone of voice, input text, etc., to determine the user's emotional state.

[0231] "Means for user viewing" refers to a means by which a user views the generated video through a terminal, and the emotion engine analyzes the user's reactions in real time while viewing and dynamically changes the video content.

[0232] This invention is a system that uses a generative AI model to recreate past TV programs and combines it with an emotion engine that recognizes the user's emotions. This system consists of three main elements: a server, a terminal, and a user.

[0233] The server is responsible for data collection, data preprocessing, training of the generative AI model, video generation, and execution of the emotion engine. Specifically, the server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, archive facilities, etc. Data collection uses Python libraries such as Requests and BeautifulSoup. The collected data is preprocessed to remove noise and unnecessary parts, and the audio data is converted to text using the Google Speech-to-Text API. This preprocessing prepares the data in a format suitable for the generative AI model.

[0234] The server then uses machine learning frameworks such as TensorFlow or PyTorch to train a generative AI model using the preprocessed data, which is then used to fill in gaps and generate a continuous video based on the user's request.

[0235] The user inputs information about the program they want to recreate (title, broadcast year, cast, etc.) through the device interface. This information is converted into JSON format and sent to the server. The server receives the user's request and searches for the corresponding program data in its database. It generates video using a generative AI model and converts it into a viewable format (e.g., MP4) using FFmpeg.

[0236] Additionally, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. The emotion engine can use OpenCV or TensorFlow, and this information is used to adjust the parameters of the generative AI model and taken into account when generating the video.

[0237] Users watch the generated video through their device. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. Users also have the option to stream or download the video.

[0238] As a concrete example, when a user requests a recreation of a 1970s variety show, an example prompt might be, "Please recreate the 1970s variety show 'XX Show.' The main character should be XX, and the atmosphere should be as joyful as possible." If the emotion engine determines that the user's emotional state is "joy," the generative AI model applies a filter that emphasizes happy scenes and bright tones. If the user is in a "sad" emotional state, the content of the video and the tone of the music are adjusted to match that emotion. In this way, a customized video experience is provided according to the user's emotions.

[0239] To implement the invention, specific hardware and software must be selected and appropriate parameters must be adjusted, but highly accurate reproduction and customization can be achieved by applying the processing steps and tools described above.

[0240] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0241] Step 1:

[0242] The server collects the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archives. Specifically, the server uses Python Requests and BeautifulSoup to perform web scraping and obtain data from APIs. The server receives a URL or API endpoint as input and obtains the collected data as output.

[0243] Step 2:

[0244] The server preprocesses the collected data. First, it converts the audio data into text using the Google Speech-to-Text API. Next, it compares the resulting text with a script to identify missing parts. It also uses regular expressions to remove noise and unnecessary characters. The collected data is given as input, and preprocessed data is obtained as output. Specifically, it performs processes such as noise removal, removal of unnecessary parts, and conversion of audio data into text.

[0245] Step 3:

[0246] The server uses the preprocessed data to train a generative AI model. It uses TensorFlow or PyTorch to build the model and sets hyperparameters (e.g., learning rate, number of epochs) to train it. The preprocessed data is given as input, and a trained generative AI model is obtained as output. Specifically, data loading, model definition, training, and validation are performed.

[0247] Step 4:

[0248] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The input information is converted to JSON format and sent to the server. The user's program information is given as input, and JSON format data is obtained as output. Specifically, form input, JSON conversion, and data transmission to the server are performed.

[0249] Step 5:

[0250] The server receives a user request and searches for the corresponding program data in a database. It uses a generative AI model to fill in missing parts and generate a continuous video. The generated video is converted into a viewable format (e.g., MP4) using FFmpeg. The user request is given as input, and the generated video file is obtained as output. Specifically, a database query, video generation, and encoding are performed.

[0251] Step 6:

[0252] The emotion engine built into the server analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. It receives real-time video feeds, audio data, and text data as input, and outputs the user's emotional state. Specifically, it uses OpenCV and TensorFlow for facial recognition, voice analysis, and text analysis.

[0253] Step 7:

[0254] The user watches the generated video through their device. The emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. The generated video and the user's reactions are given as input, and a customized viewing experience is output. Specifically, streaming playback, emotion analysis, and video content adjustment are performed.

[0255] (Application example 2)

[0256] 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."

[0257] When recreating past TV programs, it is difficult to customize the visual experience to match the viewer's emotions. Furthermore, to dynamically adjust the visual experience, it is necessary to recognize the user's emotions in real time and adjust the output of the generative AI model accordingly. Such a system does not currently exist, making it difficult to optimize the user's visual experience based on their individual emotional state.

[0258] The specific processing by the specific 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 a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, a means for dynamically adjusting the visual experience, and a means for user viewing. This makes it possible to customize the visual experience according to the user's emotions.

[0259] "Means of collecting data" refers to devices or systems for obtaining necessary data such as video, audio, scripts, etc. from the Internet, digital archives, etc.

[0260] "Preprocessing means" refers to devices and algorithms that analyze and filter collected data, remove noise and unnecessary parts, and provide high-quality data to the generative AI model.

[0261] A "means for training a generative AI model" is a device or algorithm that trains an artificial intelligence using preprocessed data, generating a model and adjusting hyperparameters.

[0262] The "means for user input" refers to an interface or device that allows the user to input information about the program that the user wants to reproduce, and that transmits the input information to the server.

[0263] "Means for generating video" refers to devices or algorithms that use generative AI models to fill in missing parts and create continuous video based on a user's request.

[0264] A "means for recognizing user emotions" is a device or algorithm that analyzes a user's facial expressions, tone of voice, and input text to determine the user's emotional state.

[0265] A "means for dynamically adjusting the visual experience" is a device or algorithm that adjusts the content and tone of a video in real time based on the user's emotions to provide the user with an optimal visual experience.

[0266] "Means for user viewing" refers to a terminal or system that allows a user to view the generated video, and that provides streaming playback or download options.

[0267] This invention is a system that reproduces past television programs and dynamically customizes the visual experience by recognizing the user's emotions. This system includes means for collecting various data, preprocessing it, and training a generative AI model, means for the user to provide input information, means for generating images, means for recognizing the user's emotions, means for dynamically adjusting the visual experience, and means for the user to view.

[0268] The server first collects the necessary data, such as video, audio, and scripts, from the internet or digital archives using APIs and scraping techniques. The collected data is then analyzed and filtered in a pre-processing stage to remove noise and unnecessary parts.

[0269] Next, the data is prepared for training a generative AI model using frameworks such as TensorFlow or PyTorch, including tuning hyperparameters. The preprocessed data is used to train the model, which then builds a generative AI model that fills in the gaps.

[0270] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). This information is converted to JSON format on the device and sent to the server. The server searches the database based on the information received from the user and retrieves the corresponding program data.

[0271] During the video generation stage, the server uses a generative AI model to fill in missing parts and create a continuous video. The generated video is then encoded and converted into a viewable format (e.g., MP4). At this time, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine their emotional state. Emotional information is fed back to the generative AI model, which dynamically adjusts the video generation parameters.

[0272] Users view the generated video using a viewing device such as a smartphone or head-mounted display. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be adjusted on the fly. The device offers streaming playback and download options to enhance the viewing experience.

[0273] As a concrete example, consider a user requesting a re-enactment of the 1970s variety show "Owarai Club." If the user smiles while watching, the emotion engine will recognize this and add fun moments and light-hearted jokes to the generated video, providing the user with a more enjoyable visual experience.

[0274] Example prompt sentence:

[0275] "Generate footage from the 1985 variety show 'Owarai Club,' and add brighter tones and fun scenes when the user is happy."

[0276] As described above, this system combines an emotion engine with a generative AI model to provide a customized visual experience that reflects the user's emotions.

[0277] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0278] Step 1: Data collection

[0279] The server collects the necessary data, such as video, audio, and scripts, from the internet or digital archives. This is done using API or scraping technology. Specifically, the server sends an API request and receives the response data in JSON format. The received data is temporarily stored in a database. A URL or API endpoint is used as input, and the collected video, audio, and script data are obtained as output.

[0280] Step 2: Data Preprocessing

[0281] The server pre-processes the collected data. This involves parsing and filtering the data to remove noise and unwanted parts. For example, it denoises the audio data and identifies missing parts of the script. The input is the collected raw data, and the output is a high-quality parsed and filtered dataset.

[0282] Step 3: Training the generative AI model

[0283] The server uses the preprocessed data to train a generative AI model. This is done using frameworks such as TensorFlow and PyTorch, and also adjusts hyperparameters. Specifically, the dataset is input into the model as training data, and parameters are optimized to fill in missing parts. The input is the preprocessed dataset, and the output is a trained generative AI model.

[0284] Step 4: Getting User Input

[0285] The terminal provides an interface for the user to input information about the program they want to replay (title, broadcast year, cast, etc.). The user inputs the information through the terminal, and the data is converted into JSON format and sent to the server. The input is the user's replay request information, and the output is the request data in JSON format.

[0286] Step 5: Image generation

[0287] The server receives a user request and searches the database for the corresponding program data. It then uses a generative AI model to fill in the missing parts and generate a continuous video. The generated video is encoded and converted into a viewable format (e.g., MP4). The input is the user request in JSON format and the trained generative AI model, and the output is the generated video data.

[0288] Step 6: Emotion Recognition

[0289] The emotion engine analyzes the user's facial expressions, voice tone, and input text to recognize the user's emotional state. This is done by collecting data in real time using the device's camera and microphone while the user is watching and applying emotion analysis algorithms. The input is the user's real-time facial and voice data, and the output is an evaluation of the user's emotional state.

[0290] Step 7: Dynamically adjust the visual experience

[0291] The server dynamically adjusts the parameters of the generative AI model based on the emotional state data obtained from the emotion engine. Specifically, if the user is happy, adjustments are made such as adding a brighter tone or a more enjoyable scene. The input is the emotional state data and the generative AI model, and the output is the adjusted video data.

[0292] Step 8: User Viewing

[0293] The user watches the generated video through the device, which offers streaming playback and download options, allowing the user to enjoy a real-time visual experience. The input is the processed video data, and the output is the user's visual experience.

[0294] In this way, the server, the terminal, and the user can work together to provide a customized visual experience according to the user's emotions.

[0295] 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.

[0296] 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.

[0297] 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.

[0298] [Second embodiment]

[0299] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0300] 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.

[0301] 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).

[0302] 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.

[0303] 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.

[0304] 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).

[0305] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0306] 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.

[0307] 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.

[0308] 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.

[0309] 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.

[0310] 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."

[0311] This invention is a system for recreating and archiving past television programs using generative AI. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing.

[0312] Specific program processing

[0313] 1. Data Collection

[0314] The server collects the necessary data such as video, audio, scripts, etc. from the Internet, libraries, and archive facilities.

[0315] The server obtains data using APIs or scraping techniques and temporarily stores it.

[0316] 2. Data Preprocessing

[0317] The server analyzes the collected data and removes noise and unnecessary parts.

[0318] The server converts the audio data into text and matches it with the script to identify missing parts.

[0319] 3. Generative Model Training

[0320] The server uses the collected data to train a generative AI model, specifically using frameworks like TensorFlow and PyTorch.

[0321] The server evaluates the model's performance every epoch during the training process and makes readjustments as needed.

[0322] 4. User Input

[0323] The user inputs information about the program they want to reproduce (program name, broadcast year, cast, etc.) through the terminal interface.

[0324] The terminal converts the user's input into JSON format and sends it to the server.

[0325] 5. Image Generation

[0326] Based on a request from a user, the server searches the database for the corresponding program data.

[0327] The server uses a generative AI model to generate a video that fills in the missing parts.

[0328] The server encodes the generated video, converts it into a viewable format, and stores it in a database.

[0329] 6. User Viewing

[0330] The user watches the generated video through a device that offers streaming playback and download options.

[0331] When a user watches a video, the device uses a cache function to provide efficient access.

[0332] Specific examples

[0333] For example, imagine a user wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. The server then collects the relevant data and fills in the missing parts with a generative AI model. The resulting video can then be streamed or downloaded and viewed by the user. This process brings a precious past show back to life in a new form, making it available for both current and future audiences to enjoy.

[0334] In this way, the present invention realizes a system that can recreate past television programs and provide them in a viewable format. This system fully utilizes the capabilities of generative AI to fill in missing parts, maintaining the continuity and quality of the video, allowing viewers to access valuable video assets from the past.

[0335] The processing flow will be explained below.

[0336] Step 1:

[0337] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to automate the retrieval of information and efficiently extract the relevant content.

[0338] Step 2:

[0339] The server analyzes and pre-processes the collected data, specifically filtering to remove noise from the video data and clean up the audio data, and uses speech recognition technology to convert the audio data into text.

[0340] Step 3:

[0341] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[0342] Step 4:

[0343] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[0344] Step 5:

[0345] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[0346] Step 6:

[0347] The user uses the interface on their terminal to input information about the program they want to reproduce (title, broadcast year, cast, etc.), which sends a specific request tailored to the user's needs to the system.

[0348] Step 7:

[0349] The device converts the request entered by the user into JSON format and sends it to the server. The server analyzes the received request and searches the database for the corresponding program data.

[0350] Step 8:

[0351] The server uses a generative AI model to fill in the gaps and generate the complete video data, using technology to generate consecutive scenes in a consistent manner while maintaining synchronization between audio and video.

[0352] Step 9:

[0353] The server encodes the generated video into a viewable format (e.g., MP4) and stores it in a database, where it can be instantly streamed or downloaded.

[0354] Step 10:

[0355] Users view the generated video through a device that offers streaming playback and download options, allowing users to freely access the video.

[0356] Step 11:

[0357] When a user watches a video, the device uses a cache function to shorten the video loading time and provide smooth playback. Any user feedback is sent to the server to help improve the system.

[0358] In this way, the server, terminal, and user play their respective roles at each step, realizing a system for reproducing past television programs.

[0359] Example 1

[0360] 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."

[0361] Many past television programs have been lost over the years, resulting in the loss of valuable cultural assets for present and future viewers. Restoring lost programs using conventional technology has been difficult and often requires advanced expertise and equipment. As a result, there is a problem in that opportunities for the general public to easily watch these programs are limited.

[0362] 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.

[0363] In this invention, the server includes a means for collecting data, a means for analyzing and filtering the data, and a means for training a generative AI model, which makes it possible to complement missing parts of past TV programs and reproduce them in a viewable form.

[0364] "Means of collecting data" refers to the means of collecting the necessary data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[0365] "Means for analyzing and filtering data" refers to means for removing noise and unnecessary parts from collected data, converting voice data into text, and matching it with a script.

[0366] "Means for training a generative AI model" means means for utilizing collected and preprocessed data to train a generative AI model and obtain highly accurate output.

[0367] "Means for users to input information" refers to an interface for users to input information about the program they wish to reproduce (e.g., program name, broadcast year, cast, etc.) and means for processing the input content.

[0368] "Means for generating images" refers to means for generating images that fill in missing parts using a pre-processed and trained generative AI model.

[0369] "Means for storing the generated video and converting it into a viewable format" refers to means for encoding the video created by the generative AI model, converting it into a common viewing format, and storing it.

[0370] The "means for user viewing" refers to an interface and system that allows a user to view the generated video by streaming playback or downloading it.

[0371] This invention is a system for recreating and archiving past television programs using generative AI. This system includes means for collecting data, means for preprocessing, means for training a generative AI model, means for users to input information, means for generating video, means for saving the generated video and converting it into a viewable format, and means for users to view it.

[0372] The server first uses a data collection method to collect the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archive facilities. Specifically, it obtains the data using APIs (e.g., YouTube Data API or open databases) or scraping technologies (e.g., Scrapy or Selenium), and temporarily stores them.

[0373] The server then uses data analysis and filtering techniques to remove noise and unwanted parts from the collected data. At this stage, FFmpeg and OpenCV are used to process the video data, and noise reduction techniques are used to process the audio data. The audio data is also converted to text using the Google Cloud Speech-to-Text API, which is then matched against a script to identify missing parts.

[0374] The server then uses a method to train the generative AI model using the preprocessed dataset. Specifically, it uses TensorFlow or PyTorch to train the model and obtain highly accurate output. During the training process, the model's performance is evaluated after each epoch, and techniques such as early stopping are applied to prevent overfitting.

[0375] The user inputs the information of the program they want to reproduce through the terminal interface using the information input means. The input content is converted to JSON format and sent to the server using the HTTPS protocol. For example, the user inputs the following prompt sentence:

[0376] "Please recreate a segment from a variety show that aired in 1975. The cast is a well-known celebrity."

[0377] The server uses a video generation method to generate a continuous video based on a user request, filling in missing parts with a generative AI model. The generated video is then encoded and converted into a viewable MP4 format. Codecs used include H.264 and HEVC. The generated video is then stored in the server's database.

[0378] Finally, the user uses a viewing means to view the generated video on a device. The device provides streaming playback or download options and utilizes a streaming server to efficiently deliver the generated video. It also uses a streaming cache function while the user is viewing to provide a smooth viewing experience.

[0379] In this way, the present invention realizes a system that can reproduce past television programs and provide them in a viewable format, thereby enabling viewers to easily access valuable video assets from the past.

[0380] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0381] Program processing steps

[0382] Step 1: Data collection

[0383] Description: The server collects the necessary data (video, audio, scripts, etc.) from the internet, libraries, and archives, using APIs (e.g., open databases) and scraping techniques (e.g., Scrapy and Selenium).

[0384] Input: Raw data from the internet, libraries and archival facilities.

[0385] Data processing / data calculation: Data acquisition via API or scraping.

[0386] Output: Captured video data, audio data, and script data.

[0387] Step 2: Data Preprocessing

[0388] Description: The server analyzes the collected data and removes noise and unnecessary parts. Specifically, it uses FFmpeg and OpenCV for video data and applies noise reduction technology to audio data. It also converts audio data to text using the Google Cloud Speech-to-Text API and matches it with the script to identify missing parts.

[0389] Input: Captured video data, audio data, and script data.

[0390] Data processing / data calculation: noise removal, speech-to-text conversion, text-to-script matching.

[0391] Output: Preprocessed video data, text data, and missing part information.

[0392] Step 3: Generative model training

[0393] Description: The server trains a generative AI model using preprocessed data. Specifically, it uses TensorFlow or PyTorch to split the dataset into training data and validation data and train the model. It evaluates the model's performance every epoch and applies techniques such as early stopping to prevent overfitting.

[0394] Input: Preprocessed video data, text data, and missing part information.

[0395] Data processing / data computation: model training, performance evaluation, and hyperparameter tuning.

[0396] Output: A trained generative AI model.

[0397] Step 4: User Input

[0398] Description: The user inputs the information of the program they want to reproduce through the device interface. The device converts the user's input into JSON format and sends it to the server using the HTTPS protocol.

[0399] Input: Program information entered by the user (e.g., program name, broadcast year, cast).

[0400] Data processing / data calculation: Convert user input into JSON format and send to the server.

[0401] Output: The request data sent to the server.

[0402] Step 5: Image generation

[0403] Description: The server searches the database for the relevant program data based on the user's request, and generates a video that fills in the missing parts using a generative AI model.The generated video is then encoded using a codec such as H.264 or HEVC, converted to MP4 format, and saved in the database.

[0404] Input: User request data, trained generative AI model, and program data from the database.

[0405] Data processing / data calculation: database search, generation of complementary video, video encoding, conversion to MP4 format.

[0406] Output: The generated video data.

[0407] Step 6: User Viewing

[0408] Description: The user uses a device to watch the generated video. The device uses a streaming server (e.g., NGINX or Wowza) to stream the video or provides a download option. During viewing, a cache function is used to provide a smooth viewing experience.

[0409] Input: Generated video data.

[0410] Data processing / data calculation: streaming distribution, cache processing.

[0411] Output: The video that the user can see.

[0412] (Application example 1)

[0413] 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."

[0414] There is a need to recreate valuable past television programs and video assets so that they can be viewed by current and future viewers. However, over the years, the original video data can deteriorate or become missing. Furthermore, when recreating past video, noise and unnecessary information are often mixed in, significantly reducing viewing quality. Therefore, a system is needed that can efficiently recreate past television programs and provide them to viewers with high quality.

[0415] 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.

[0416] In this invention, the server includes means for collecting data, means for preprocessing, means for training a generative AI model, means for user input, means for generating video, means for user viewing, means for analyzing the collected data and processing it to reproduce video of a specified program, means for providing an interface for users to request past video, and means for converting video generated using hardware into a viewable format, thereby making it possible to reproduce past TV programs in high quality and efficiently provide video requested by viewers.

[0417] "Data collection means" refers to the devices and methods used to collect the necessary data from the Internet and archive facilities.

[0418] "Preprocessing means" refers to a device or method that analyzes collected data and removes noise and unnecessary parts.

[0419] A "means for training a generative AI model" is a device or method that uses collected data to train an AI model and acquire knowledge to fill in any gaps.

[0420] "User input means" refers to a device or method for a user to input information about a program that the user wants to reproduce (such as the program name, broadcast year, and cast members).

[0421] "Image generation means" refers to a device or method that uses preprocessed data and a generative AI model to generate images that fill in missing parts.

[0422] "User viewing means" refers to a device or method that allows a user to view the generated video.

[0423] The "collected data analysis means" refers to a device or method that analyzes the collected data and performs the processing necessary to reproduce the video of the specified program.

[0424] The "user interface providing means" refers to a device or method that provides an interface for a user to request past video.

[0425] "Video conversion means" refers to a device or method that converts video generated using hardware into a viewable format.

[0426] This invention is a system that uses generative AI to recreate and archive past television programs. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing. Specific embodiments of these means are described below.

[0427] First, in the data collection process, the server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. This is achieved using APIs and scraping technology. The collected data is temporarily stored and used for subsequent processing.

[0428] Next, the pre-processing means analyzes the data collected by the server and removes noise and unnecessary parts. Specifically, it performs frame analysis of the collected video data and converts audio data into text. This converts the data into a format suitable for video complementation.

[0429] The generative AI model training method uses the collected data to train the generative AI model. Using frameworks like TensorFlow or PyTorch, the model's performance is evaluated every epoch and retuned as needed. As the model learns, it acquires knowledge to fill in missing parts of the video.

[0430] The user input means allows the user to input information about the program they want to reproduce (such as the program name, broadcast year, and cast members) through the device's interface. The device converts the user's input into JSON format and sends it to the server, allowing the server to efficiently search for specific program data.

[0431] In the video generation method, the server uses the generative AI model to generate videos that fill in missing parts. These videos are encoded into a format that can be viewed by users and stored in a database.

[0432] Finally, as a means for user viewing, users can view the generated video through their devices, with streaming playback and download options available, and a cache function for efficient access.

[0433] As a concrete example, consider a user who wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. In this case, the prompt sentence is "Please recreate the 1970 variety show 'Retro Variety Show'." The server collects the relevant data and fills in the missing parts with a generative AI model. As a result, the user can stream or download the generated video and watch it.

[0434] As described above, the present invention makes it possible to reproduce past television programs with high quality and efficiently provide the video requested by the viewer.

[0435] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0436] Step 1: Data collection

[0437] The server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. Specifically, it uses APIs and scraping technology to obtain data related to the specified program. The input is the program name and broadcast year, and the output is temporarily stored video, audio, and text data.

[0438] Step 2: Data Preprocessing

[0439] The server analyzes the collected data and processes it to remove noise and unnecessary parts. Specifically, it uses OpenCV to analyze frames of video data and applies a noise removal filter. Additionally, audio data is converted into text using speech recognition technology and matched with a script. The input is the raw collected data, and the output is the clean data that has been analyzed and filtered.

[0440] Step 3: Training the generative AI model

[0441] The server uses the preprocessed data to train a generative AI model. This uses frameworks such as TensorFlow and PyTorch. Specifically, it pairs video and text data to create a training set, and then trains the model based on that. The input is the preprocessed data, and the output is a fully trained generative AI model.

[0442] Step 4: User Input

[0443] The user inputs information about the program they want to play (such as the program name, broadcast year, and cast) through the device's interface. The device converts this input into JSON format and sends it to the server. The input is the user's program request information, and the output is JSON format data.

[0444] Step 5: Image generation

[0445] Based on a user request, the server uses a generative AI model to generate a video with missing parts filled in. Specifically, it uses a trained model to generate reconstructed video frames from clean data. The input is the user request and clean data, and the output is the filled video data.

[0446] Step 6: User Viewing

[0447] The user watches the generated video through a device. The device provides a player for streaming the video, and also offers a download option. The input is the generated video data, and the output is a viewable video.

[0448] As described above, by performing appropriate data processing and data calculations based on the input data at each step, a system is realized that reproduces past television programs with high quality and provides the video requested by the user.

[0449] 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.

[0450] This invention is a system that uses generative AI to recreate past television programs and combines it with an emotion engine that recognizes user emotions. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for user viewing, and an emotion engine that recognizes user emotions.

[0451] Specific program processing

[0452] 1. Data Collection

[0453] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives using APIs and scraping technology.

[0454] 2. Data Preprocessing

[0455] The server analyzes the collected data, filtering out noise and unnecessary parts, converts the audio data to text, and matches it with a script to identify missing parts, thereby preparing the data for subsequent input into a generative AI model.

[0456] 3. Generative Model Training

[0457] The server trains a generative AI model using the preprocessed data. Using the prepared dataset, it builds a generative AI model that fills in the gaps. The training uses the TensorFlow and PyTorch frameworks, and also adjusts hyperparameters.

[0458] 4. User Input

[0459] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The device converts the input information into JSON format and sends it to the server.

[0460] 5. Image Generation

[0461] The server receives the user's request, searches the database for the corresponding program data, uses a generative AI model to fill in the gaps and create a continuous video, and then encodes the video and converts it into a viewable format (e.g., MP4).

[0462] 6. Operation of the Emotion Engine

[0463] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. This information is taken into account when generating the video, and the parameters of the generative AI model are adjusted as needed.

[0464] 7. User Viewing

[0465] Users watch the generated video through the device. An emotion engine analyzes users' reactions in real time, and the content and tone of the video may be dynamically changed. The device offers streaming and download options to enhance the viewing experience.

[0466] Specific examples

[0467] For example, consider a user requesting a re-enactment of a 1970s variety show. If the emotion engine determines that the user is in a positive emotional state (e.g., joy), the generative AI model applies a filter that emphasizes more joyful scenes and brighter tones. On the other hand, if the user is in a sad emotional state, the video content and musical tone are adjusted to match that emotion. In this way, a customized video experience is provided based on the user's emotions.

[0468] This system not only makes precious programs from the past available to current and future audiences in new formats, but also customizes the visual experience based on the user's emotional state. The collaboration between generative AI and the emotion engine makes it possible to provide even more compelling visual content.

[0469] The processing flow will be explained below.

[0470] Step 1:

[0471] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to efficiently extract the relevant content.

[0472] Step 2:

[0473] The server analyzes the collected data and removes noise and unnecessary parts, specifically by performing noise reduction processing on video data and filtering on audio data, and also by using voice recognition technology to convert audio data into text.

[0474] Step 3:

[0475] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[0476] Step 4:

[0477] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[0478] Step 5:

[0479] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[0480] Step 6:

[0481] The user uses the interface on their device to input information about the program they want to replay (title, broadcast year, cast, etc.), and the system processes the request based on this information.

[0482] Step 7:

[0483] The device converts the request entered by the user into JSON format and sends it to the server, which then analyzes the request and searches the database for the corresponding program data.

[0484] Step 8:

[0485] The server uses a generative AI model to fill in the gaps and generate the complete video data, generating consecutive scenes in a consistent manner while maintaining audio and video synchronization.

[0486] Step 9:

[0487] The server encodes the generated video into a viewable format (e.g., MP4 format) and stores it in a database. The stored video can then be used for streaming or download.

[0488] Step 10:

[0489] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state, and the analysis results are sent to the server when the video is generated.

[0490] Step 11:

[0491] The server receives data from the emotion engine and adjusts the parameters of the generative AI model based on the user's emotional state, generating customized video according to the user's emotions.

[0492] Step 12:

[0493] The user watches the generated video through a device that provides streaming and download options to allow the user to watch the video in the appropriate way.

[0494] Step 13:

[0495] As users watch videos, the emotion engine analyzes their reactions in real time and transmits the information to the server, which can then dynamically adjust the content and tone of the video to provide the best possible viewing experience based on the user's emotions.

[0496] Example 2

[0497] 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."

[0498] In systems that reproduce past TV programs, there is a need for a method that not only plays back existing data but also provides a customized experience based on the user's emotions. Another challenge is to efficiently perform preprocessing such as noise removal from collected data and text conversion of audio data to achieve high-precision video generation.

[0499] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, and a means for user viewing. This makes it possible to reproduce past TV programs with high accuracy and dynamically change the video content and tone according to the user's emotions.

[0500] "Means of collecting data" refers to the means of obtaining data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[0501] "Preprocessing means" refers to the means for analyzing collected data, removing noise, deleting unnecessary parts, and preparing the data in a format suitable for a generative AI model.

[0502] "Means for training a generative AI model" refers to a means for training an AI model using preprocessed data and building a highly accurate model that can fill in missing parts.

[0503] The "means for user input" refers to a means by which a user inputs information about a program that the user wants to reproduce through a terminal and transmits the information to the server.

[0504] "Means for generating video" refers to a means for using a generative AI model to create video of a program requested by a user and to fill in missing parts to generate continuous video.

[0505] The "means for recognizing the user's emotions" is a means for analyzing the user's facial expressions, tone of voice, input text, etc., to determine the user's emotional state.

[0506] "Means for user viewing" refers to a means by which a user views the generated video through a terminal, and the emotion engine analyzes the user's reactions in real time while viewing and dynamically changes the video content.

[0507] This invention is a system that uses a generative AI model to recreate past TV programs and combines it with an emotion engine that recognizes the user's emotions. This system consists of three main elements: a server, a terminal, and a user.

[0508] The server is responsible for data collection, data preprocessing, training of the generative AI model, video generation, and execution of the emotion engine. Specifically, the server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, archive facilities, etc. Data collection uses Python libraries such as Requests and BeautifulSoup. The collected data is preprocessed to remove noise and unnecessary parts, and the audio data is converted to text using the Google Speech-to-Text API. This preprocessing prepares the data in a format suitable for the generative AI model.

[0509] The server then uses machine learning frameworks such as TensorFlow or PyTorch to train a generative AI model using the preprocessed data, which is then used to fill in gaps and generate a continuous video based on the user's request.

[0510] The user inputs information about the program they want to recreate (title, broadcast year, cast, etc.) through the device interface. This information is converted into JSON format and sent to the server. The server receives the user's request and searches for the corresponding program data in its database. It generates video using a generative AI model and converts it into a viewable format (e.g., MP4) using FFmpeg.

[0511] Additionally, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. The emotion engine can use OpenCV or TensorFlow, and this information is used to adjust the parameters of the generative AI model and taken into account when generating the video.

[0512] Users watch the generated video through their device. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. Users also have the option to stream or download the video.

[0513] As a concrete example, when a user requests a recreation of a 1970s variety show, an example prompt might be, "Please recreate the 1970s variety show 'XX Show.' The main character should be XX, and the atmosphere should be as joyful as possible." If the emotion engine determines that the user's emotional state is "joy," the generative AI model applies a filter that emphasizes happy scenes and bright tones. If the user is in a "sad" emotional state, the content of the video and the tone of the music are adjusted to match that emotion. In this way, a customized video experience is provided according to the user's emotions.

[0514] To implement the invention, specific hardware and software must be selected and appropriate parameters must be adjusted, but highly accurate reproduction and customization can be achieved by applying the processing steps and tools described above.

[0515] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0516] Step 1:

[0517] The server collects the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archives. Specifically, the server uses Python Requests and BeautifulSoup to perform web scraping and obtain data from APIs. The server receives a URL or API endpoint as input and obtains the collected data as output.

[0518] Step 2:

[0519] The server preprocesses the collected data. First, it converts the audio data into text using the Google Speech-to-Text API. Next, it compares the resulting text with a script to identify missing parts. It also uses regular expressions to remove noise and unnecessary characters. The collected data is given as input, and preprocessed data is obtained as output. Specifically, it performs processes such as noise removal, removal of unnecessary parts, and conversion of audio data into text.

[0520] Step 3:

[0521] The server uses the preprocessed data to train a generative AI model. It uses TensorFlow or PyTorch to build the model and sets hyperparameters (e.g., learning rate, number of epochs) to train it. The preprocessed data is given as input, and a trained generative AI model is obtained as output. Specifically, data loading, model definition, training, and validation are performed.

[0522] Step 4:

[0523] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The input information is converted to JSON format and sent to the server. The user's program information is given as input, and JSON format data is obtained as output. Specifically, form input, JSON conversion, and data transmission to the server are performed.

[0524] Step 5:

[0525] The server receives a user request and searches for the corresponding program data in a database. It uses a generative AI model to fill in missing parts and generate a continuous video. The generated video is converted into a viewable format (e.g., MP4) using FFmpeg. The user request is given as input, and the generated video file is obtained as output. Specifically, a database query, video generation, and encoding are performed.

[0526] Step 6:

[0527] The emotion engine built into the server analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. It receives real-time video feeds, audio data, and text data as input, and outputs the user's emotional state. Specifically, it uses OpenCV and TensorFlow for facial recognition, voice analysis, and text analysis.

[0528] Step 7:

[0529] The user watches the generated video through their device. The emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. The generated video and the user's reactions are given as input, and a customized viewing experience is output. Specifically, streaming playback, emotion analysis, and video content adjustment are performed.

[0530] (Application example 2)

[0531] 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."

[0532] When recreating past TV programs, it is difficult to customize the visual experience to match the viewer's emotions. Furthermore, to dynamically adjust the visual experience, it is necessary to recognize the user's emotions in real time and adjust the output of the generative AI model accordingly. Such a system does not currently exist, making it difficult to optimize the user's visual experience based on their individual emotional state.

[0533] The specific processing by the specific 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 a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, a means for dynamically adjusting the visual experience, and a means for user viewing. This makes it possible to customize the visual experience according to the user's emotions.

[0534] "Means of collecting data" refers to devices or systems for obtaining necessary data such as video, audio, scripts, etc. from the Internet, digital archives, etc.

[0535] "Preprocessing means" refers to devices and algorithms that analyze and filter collected data, remove noise and unnecessary parts, and provide high-quality data to the generative AI model.

[0536] A "means for training a generative AI model" is a device or algorithm that trains an artificial intelligence using preprocessed data, generating a model and adjusting hyperparameters.

[0537] The "means for user input" refers to an interface or device that allows the user to input information about the program that the user wants to reproduce, and that transmits the input information to the server.

[0538] "Means for generating video" refers to devices or algorithms that use generative AI models to fill in missing parts and create continuous video based on a user's request.

[0539] A "means for recognizing user emotions" is a device or algorithm that analyzes a user's facial expressions, tone of voice, and input text to determine the user's emotional state.

[0540] A "means for dynamically adjusting the visual experience" is a device or algorithm that adjusts the content and tone of a video in real time based on the user's emotions to provide the user with an optimal visual experience.

[0541] "Means for user viewing" refers to a terminal or system that allows a user to view the generated video, and that provides streaming playback or download options.

[0542] This invention is a system that reproduces past television programs and dynamically customizes the visual experience by recognizing the user's emotions. This system includes means for collecting various data, preprocessing it, and training a generative AI model, means for the user to provide input information, means for generating images, means for recognizing the user's emotions, means for dynamically adjusting the visual experience, and means for the user to view.

[0543] The server first collects the necessary data, such as video, audio, and scripts, from the internet or digital archives using APIs and scraping techniques. The collected data is then analyzed and filtered in a pre-processing stage to remove noise and unnecessary parts.

[0544] Next, the data is prepared for training a generative AI model using frameworks such as TensorFlow or PyTorch, including tuning hyperparameters. The preprocessed data is used to train the model, which then builds a generative AI model that fills in the gaps.

[0545] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). This information is converted to JSON format on the device and sent to the server. The server searches the database based on the information received from the user and retrieves the corresponding program data.

[0546] During the video generation stage, the server uses a generative AI model to fill in missing parts and create a continuous video. The generated video is then encoded and converted into a viewable format (e.g., MP4). At this time, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine their emotional state. Emotional information is fed back to the generative AI model, which dynamically adjusts the video generation parameters.

[0547] Users view the generated video using a viewing device such as a smartphone or head-mounted display. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be adjusted on the fly. The device offers streaming playback and download options to enhance the viewing experience.

[0548] As a concrete example, consider a user requesting a re-enactment of the 1970s variety show "Owarai Club." If the user smiles while watching, the emotion engine will recognize this and add fun moments and light-hearted jokes to the generated video, providing the user with a more enjoyable visual experience.

[0549] Example prompt sentence:

[0550] "Generate footage from the 1985 variety show 'Owarai Club,' and add brighter tones and fun scenes when the user is happy."

[0551] As described above, this system combines an emotion engine with a generative AI model to provide a customized visual experience that reflects the user's emotions.

[0552] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0553] Step 1: Data collection

[0554] The server collects the necessary data, such as video, audio, and scripts, from the internet or digital archives. This is done using API or scraping technology. Specifically, the server sends an API request and receives the response data in JSON format. The received data is temporarily stored in a database. A URL or API endpoint is used as input, and the collected video, audio, and script data are obtained as output.

[0555] Step 2: Data Preprocessing

[0556] The server pre-processes the collected data. This involves parsing and filtering the data to remove noise and unwanted parts. For example, it denoises the audio data and identifies missing parts of the script. The input is the collected raw data, and the output is a high-quality parsed and filtered dataset.

[0557] Step 3: Training the generative AI model

[0558] The server uses the preprocessed data to train a generative AI model. This is done using frameworks such as TensorFlow and PyTorch, and also adjusts hyperparameters. Specifically, the dataset is input into the model as training data, and parameters are optimized to fill in missing parts. The input is the preprocessed dataset, and the output is a trained generative AI model.

[0559] Step 4: Getting User Input

[0560] The terminal provides an interface for the user to input information about the program they want to replay (title, broadcast year, cast, etc.). The user inputs the information through the terminal, and the data is converted into JSON format and sent to the server. The input is the user's replay request information, and the output is the request data in JSON format.

[0561] Step 5: Image generation

[0562] The server receives a user request and searches the database for the corresponding program data. It then uses a generative AI model to fill in the missing parts and generate a continuous video. The generated video is encoded and converted into a viewable format (e.g., MP4). The input is the user request in JSON format and the trained generative AI model, and the output is the generated video data.

[0563] Step 6: Emotion Recognition

[0564] The emotion engine analyzes the user's facial expressions, voice tone, and input text to recognize the user's emotional state. This is done by collecting data in real time using the device's camera and microphone while the user is watching and applying emotion analysis algorithms. The input is the user's real-time facial and voice data, and the output is an evaluation of the user's emotional state.

[0565] Step 7: Dynamically adjust the visual experience

[0566] The server dynamically adjusts the parameters of the generative AI model based on the emotional state data obtained from the emotion engine. Specifically, if the user is happy, adjustments are made such as adding a brighter tone or a more enjoyable scene. The input is the emotional state data and the generative AI model, and the output is the adjusted video data.

[0567] Step 8: User Viewing

[0568] The user watches the generated video through the device, which offers streaming playback and download options, allowing the user to enjoy a real-time visual experience. The input is the processed video data, and the output is the user's visual experience.

[0569] In this way, the server, the terminal, and the user can work together to provide a customized visual experience according to the user's emotions.

[0570] 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.

[0571] 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.

[0572] 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.

[0573] [Third embodiment]

[0574] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0575] 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.

[0576] 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).

[0577] 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.

[0578] 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.

[0579] 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).

[0580] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0581] 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.

[0582] 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.

[0583] 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.

[0584] 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.

[0585] 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."

[0586] This invention is a system for recreating and archiving past television programs using generative AI. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing.

[0587] Specific program processing

[0588] 1. Data Collection

[0589] The server collects the necessary data such as video, audio, scripts, etc. from the Internet, libraries, and archive facilities.

[0590] The server obtains data using APIs or scraping techniques and temporarily stores it.

[0591] 2. Data Preprocessing

[0592] The server analyzes the collected data and removes noise and unnecessary parts.

[0593] The server converts the audio data into text and matches it with the script to identify missing parts.

[0594] 3. Generative Model Training

[0595] The server uses the collected data to train a generative AI model, specifically using frameworks like TensorFlow and PyTorch.

[0596] The server evaluates the model's performance every epoch during the training process and makes readjustments as needed.

[0597] 4. User Input

[0598] The user inputs information about the program they want to reproduce (program name, broadcast year, cast, etc.) through the terminal interface.

[0599] The terminal converts the user's input into JSON format and sends it to the server.

[0600] 5. Image Generation

[0601] Based on a request from a user, the server searches the database for the corresponding program data.

[0602] The server uses a generative AI model to generate a video that fills in the missing parts.

[0603] The server encodes the generated video, converts it into a viewable format, and stores it in a database.

[0604] 6. User Viewing

[0605] The user watches the generated video through a device that offers streaming playback and download options.

[0606] When a user watches a video, the device uses a cache function to provide efficient access.

[0607] Specific examples

[0608] For example, imagine a user wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. The server then collects the relevant data and fills in the missing parts with a generative AI model. The resulting video can then be streamed or downloaded and viewed by the user. This process brings a precious past show back to life in a new form, making it available for both current and future audiences to enjoy.

[0609] In this way, the present invention realizes a system that can recreate past television programs and provide them in a viewable format. This system fully utilizes the capabilities of generative AI to fill in missing parts, maintaining the continuity and quality of the video, allowing viewers to access valuable video assets from the past.

[0610] The processing flow will be explained below.

[0611] Step 1:

[0612] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to automate the retrieval of information and efficiently extract the relevant content.

[0613] Step 2:

[0614] The server analyzes and pre-processes the collected data, specifically filtering to remove noise from the video data and clean up the audio data, and uses speech recognition technology to convert the audio data into text.

[0615] Step 3:

[0616] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[0617] Step 4:

[0618] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[0619] Step 5:

[0620] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[0621] Step 6:

[0622] The user uses the interface on their terminal to input information about the program they want to reproduce (title, broadcast year, cast, etc.), which sends a specific request tailored to the user's needs to the system.

[0623] Step 7:

[0624] The device converts the request entered by the user into JSON format and sends it to the server. The server analyzes the received request and searches the database for the corresponding program data.

[0625] Step 8:

[0626] The server uses a generative AI model to fill in the gaps and generate the complete video data, using technology to generate consecutive scenes in a consistent manner while maintaining synchronization between audio and video.

[0627] Step 9:

[0628] The server encodes the generated video into a viewable format (e.g., MP4) and stores it in a database, where it can be instantly streamed or downloaded.

[0629] Step 10:

[0630] Users view the generated video through a device that offers streaming playback and download options, allowing users to freely access the video.

[0631] Step 11:

[0632] When a user watches a video, the device uses a cache function to shorten the video loading time and provide smooth playback. Any user feedback is sent to the server to help improve the system.

[0633] In this way, the server, terminal, and user play their respective roles at each step, realizing a system for reproducing past television programs.

[0634] Example 1

[0635] 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."

[0636] Many past television programs have been lost over the years, resulting in the loss of valuable cultural assets for present and future viewers. Restoring lost programs using conventional technology has been difficult and often requires advanced expertise and equipment. As a result, there is a problem in that opportunities for the general public to easily watch these programs are limited.

[0637] 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.

[0638] In this invention, the server includes a means for collecting data, a means for analyzing and filtering the data, and a means for training a generative AI model, which makes it possible to complement missing parts of past TV programs and reproduce them in a viewable form.

[0639] "Means of collecting data" refers to the means of collecting the necessary data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[0640] "Means for analyzing and filtering data" refers to means for removing noise and unnecessary parts from collected data, converting voice data into text, and matching it with a script.

[0641] "Means for training a generative AI model" means means for utilizing collected and preprocessed data to train a generative AI model and obtain highly accurate output.

[0642] "Means for users to input information" refers to an interface for users to input information about the program they wish to reproduce (e.g., program name, broadcast year, cast, etc.) and means for processing the input content.

[0643] "Means for generating images" refers to means for generating images that fill in missing parts using a pre-processed and trained generative AI model.

[0644] "Means for storing the generated video and converting it into a viewable format" refers to means for encoding the video created by the generative AI model, converting it into a common viewing format, and storing it.

[0645] The "means for user viewing" refers to an interface and system that allows a user to view the generated video by streaming playback or downloading it.

[0646] This invention is a system for recreating and archiving past television programs using generative AI. This system includes means for collecting data, means for preprocessing, means for training a generative AI model, means for users to input information, means for generating video, means for saving the generated video and converting it into a viewable format, and means for users to view it.

[0647] The server first uses a data collection method to collect the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archive facilities. Specifically, it obtains the data using APIs (e.g., YouTube Data API or open databases) or scraping technologies (e.g., Scrapy or Selenium), and temporarily stores them.

[0648] The server then uses data analysis and filtering techniques to remove noise and unwanted parts from the collected data. At this stage, FFmpeg and OpenCV are used to process the video data, and noise reduction techniques are used to process the audio data. The audio data is also converted to text using the Google Cloud Speech-to-Text API, which is then matched against a script to identify missing parts.

[0649] The server then uses a method to train the generative AI model using the preprocessed dataset. Specifically, it uses TensorFlow or PyTorch to train the model and obtain highly accurate output. During the training process, the model's performance is evaluated after each epoch, and techniques such as early stopping are applied to prevent overfitting.

[0650] The user inputs the information of the program they want to reproduce through the terminal interface using the information input means. The input content is converted to JSON format and sent to the server using the HTTPS protocol. For example, the user inputs the following prompt sentence:

[0651] "Please recreate a segment from a variety show that aired in 1975. The cast is a well-known celebrity."

[0652] The server uses a video generation method to generate a continuous video based on a user request, filling in missing parts with a generative AI model. The generated video is then encoded and converted into a viewable MP4 format. Codecs used include H.264 and HEVC. The generated video is then stored in the server's database.

[0653] Finally, the user uses a viewing means to view the generated video on a device. The device provides streaming playback or download options and utilizes a streaming server to efficiently deliver the generated video. It also uses a streaming cache function while the user is viewing to provide a smooth viewing experience.

[0654] In this way, the present invention realizes a system that can reproduce past television programs and provide them in a viewable format, thereby enabling viewers to easily access valuable video assets from the past.

[0655] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0656] Program processing steps

[0657] Step 1: Data collection

[0658] Description: The server collects the necessary data (video, audio, scripts, etc.) from the internet, libraries, and archives, using APIs (e.g., open databases) and scraping techniques (e.g., Scrapy and Selenium).

[0659] Input: Raw data from the internet, libraries and archival facilities.

[0660] Data processing / data calculation: Data acquisition via API or scraping.

[0661] Output: Captured video data, audio data, and script data.

[0662] Step 2: Data Preprocessing

[0663] Description: The server analyzes the collected data and removes noise and unnecessary parts. Specifically, it uses FFmpeg and OpenCV for video data and applies noise reduction technology to audio data. It also converts audio data to text using the Google Cloud Speech-to-Text API and matches it with the script to identify missing parts.

[0664] Input: Captured video data, audio data, and script data.

[0665] Data processing / data calculation: noise removal, speech-to-text conversion, text-to-script matching.

[0666] Output: Preprocessed video data, text data, and missing part information.

[0667] Step 3: Generative model training

[0668] Description: The server trains a generative AI model using preprocessed data. Specifically, it uses TensorFlow or PyTorch to split the dataset into training data and validation data and train the model. It evaluates the model's performance every epoch and applies techniques such as early stopping to prevent overfitting.

[0669] Input: Preprocessed video data, text data, and missing part information.

[0670] Data processing / data computation: model training, performance evaluation, and hyperparameter tuning.

[0671] Output: A trained generative AI model.

[0672] Step 4: User Input

[0673] Description: The user inputs the information of the program they want to reproduce through the device interface. The device converts the user's input into JSON format and sends it to the server using the HTTPS protocol.

[0674] Input: Program information entered by the user (e.g., program name, broadcast year, cast).

[0675] Data processing / data calculation: Convert user input into JSON format and send to the server.

[0676] Output: The request data sent to the server.

[0677] Step 5: Image generation

[0678] Description: The server searches the database for the relevant program data based on the user's request, and generates a video that fills in the missing parts using a generative AI model.The generated video is then encoded using a codec such as H.264 or HEVC, converted to MP4 format, and saved in the database.

[0679] Input: User request data, trained generative AI model, and program data from the database.

[0680] Data processing / data calculation: database search, generation of complementary video, video encoding, conversion to MP4 format.

[0681] Output: The generated video data.

[0682] Step 6: User Viewing

[0683] Description: The user uses a device to watch the generated video. The device uses a streaming server (e.g., NGINX or Wowza) to stream the video or provides a download option. During viewing, a cache function is used to provide a smooth viewing experience.

[0684] Input: Generated video data.

[0685] Data processing / data calculation: streaming distribution, cache processing.

[0686] Output: The video that the user can see.

[0687] (Application example 1)

[0688] 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."

[0689] There is a need to recreate valuable past television programs and video assets so that they can be viewed by current and future viewers. However, over the years, the original video data can deteriorate or become missing. Furthermore, when recreating past video, noise and unnecessary information are often mixed in, significantly reducing viewing quality. Therefore, a system is needed that can efficiently recreate past television programs and provide them to viewers with high quality.

[0690] 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.

[0691] In this invention, the server includes means for collecting data, means for preprocessing, means for training a generative AI model, means for user input, means for generating video, means for user viewing, means for analyzing the collected data and processing it to reproduce video of a specified program, means for providing an interface for users to request past video, and means for converting video generated using hardware into a viewable format, thereby making it possible to reproduce past TV programs in high quality and efficiently provide video requested by viewers.

[0692] "Data collection means" refers to the devices and methods used to collect the necessary data from the Internet and archive facilities.

[0693] "Preprocessing means" refers to a device or method that analyzes collected data and removes noise and unnecessary parts.

[0694] A "means for training a generative AI model" is a device or method that uses collected data to train an AI model and acquire knowledge to fill in any gaps.

[0695] "User input means" refers to a device or method for a user to input information about a program that the user wants to reproduce (such as the program name, broadcast year, and cast members).

[0696] "Image generation means" refers to a device or method that uses preprocessed data and a generative AI model to generate images that fill in missing parts.

[0697] "User viewing means" refers to a device or method that allows a user to view the generated video.

[0698] The "collected data analysis means" refers to a device or method that analyzes the collected data and performs the processing necessary to reproduce the video of the specified program.

[0699] The "user interface providing means" refers to a device or method that provides an interface for a user to request past video.

[0700] "Video conversion means" refers to a device or method that converts video generated using hardware into a viewable format.

[0701] This invention is a system that uses generative AI to recreate and archive past television programs. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing. Specific embodiments of these means are described below.

[0702] First, in the data collection process, the server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. This is achieved using APIs and scraping technology. The collected data is temporarily stored and used for subsequent processing.

[0703] Next, the pre-processing means analyzes the data collected by the server and removes noise and unnecessary parts. Specifically, it performs frame analysis of the collected video data and converts audio data into text. This converts the data into a format suitable for video complementation.

[0704] The generative AI model training method uses the collected data to train the generative AI model. Using frameworks like TensorFlow or PyTorch, the model's performance is evaluated every epoch and retuned as needed. As the model learns, it acquires knowledge to fill in missing parts of the video.

[0705] The user input means allows the user to input information about the program they want to reproduce (such as the program name, broadcast year, and cast members) through the device's interface. The device converts the user's input into JSON format and sends it to the server, allowing the server to efficiently search for specific program data.

[0706] In the video generation method, the server uses the generative AI model to generate videos that fill in missing parts. These videos are encoded into a format that can be viewed by users and stored in a database.

[0707] Finally, as a means for user viewing, users can view the generated video through their devices, with streaming playback and download options available, and a cache function for efficient access.

[0708] As a concrete example, consider a user who wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. In this case, the prompt sentence is "Please recreate the 1970 variety show 'Retro Variety Show'." The server collects the relevant data and fills in the missing parts with a generative AI model. As a result, the user can stream or download the generated video and watch it.

[0709] As described above, the present invention makes it possible to reproduce past television programs with high quality and efficiently provide the video requested by the viewer.

[0710] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0711] Step 1: Data collection

[0712] The server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. Specifically, it uses APIs and scraping technology to obtain data related to the specified program. The input is the program name and broadcast year, and the output is temporarily stored video, audio, and text data.

[0713] Step 2: Data Preprocessing

[0714] The server analyzes the collected data and processes it to remove noise and unnecessary parts. Specifically, it uses OpenCV to analyze frames of video data and applies a noise removal filter. Additionally, audio data is converted into text using speech recognition technology and matched with a script. The input is the raw collected data, and the output is the clean data that has been analyzed and filtered.

[0715] Step 3: Training the generative AI model

[0716] The server uses the preprocessed data to train a generative AI model. This uses frameworks such as TensorFlow and PyTorch. Specifically, it pairs video and text data to create a training set, and then trains the model based on that. The input is the preprocessed data, and the output is a fully trained generative AI model.

[0717] Step 4: User Input

[0718] The user inputs information about the program they want to play (such as the program name, broadcast year, and cast) through the device's interface. The device converts this input into JSON format and sends it to the server. The input is the user's program request information, and the output is JSON format data.

[0719] Step 5: Image generation

[0720] Based on a user request, the server uses a generative AI model to generate a video with missing parts filled in. Specifically, it uses a trained model to generate reconstructed video frames from clean data. The input is the user request and clean data, and the output is the filled video data.

[0721] Step 6: User Viewing

[0722] The user watches the generated video through a device. The device provides a player for streaming the video, and also offers a download option. The input is the generated video data, and the output is a viewable video.

[0723] As described above, by performing appropriate data processing and data calculations based on the input data at each step, a system is realized that reproduces past television programs with high quality and provides the video requested by the user.

[0724] 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.

[0725] This invention is a system that uses generative AI to recreate past television programs and combines it with an emotion engine that recognizes user emotions. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for user viewing, and an emotion engine that recognizes user emotions.

[0726] Specific program processing

[0727] 1. Data Collection

[0728] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives using APIs and scraping technology.

[0729] 2. Data Preprocessing

[0730] The server analyzes the collected data, filtering out noise and unnecessary parts, converts the audio data to text, and matches it with a script to identify missing parts, thereby preparing the data for subsequent input into a generative AI model.

[0731] 3. Generative Model Training

[0732] The server trains a generative AI model using the preprocessed data. Using the prepared dataset, it builds a generative AI model that fills in the gaps. The training uses the TensorFlow and PyTorch frameworks, and also adjusts hyperparameters.

[0733] 4. User Input

[0734] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The device converts the input information into JSON format and sends it to the server.

[0735] 5. Image Generation

[0736] The server receives the user's request, searches the database for the corresponding program data, uses a generative AI model to fill in the gaps and create a continuous video, and then encodes the video and converts it into a viewable format (e.g., MP4).

[0737] 6. Operation of the Emotion Engine

[0738] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. This information is taken into account when generating the video, and the parameters of the generative AI model are adjusted as needed.

[0739] 7. User Viewing

[0740] Users watch the generated video through the device. An emotion engine analyzes users' reactions in real time, and the content and tone of the video may be dynamically changed. The device offers streaming and download options to enhance the viewing experience.

[0741] Specific examples

[0742] For example, consider a user requesting a re-enactment of a 1970s variety show. If the emotion engine determines that the user is in a positive emotional state (e.g., joy), the generative AI model applies a filter that emphasizes more joyful scenes and brighter tones. On the other hand, if the user is in a sad emotional state, the video content and musical tone are adjusted to match that emotion. In this way, a customized video experience is provided based on the user's emotions.

[0743] This system not only makes precious programs from the past available to current and future audiences in new formats, but also customizes the visual experience based on the user's emotional state. The collaboration between generative AI and the emotion engine makes it possible to provide even more compelling visual content.

[0744] The processing flow will be explained below.

[0745] Step 1:

[0746] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to efficiently extract the relevant content.

[0747] Step 2:

[0748] The server analyzes the collected data and removes noise and unnecessary parts, specifically by performing noise reduction processing on video data and filtering on audio data, and also by using voice recognition technology to convert audio data into text.

[0749] Step 3:

[0750] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[0751] Step 4:

[0752] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[0753] Step 5:

[0754] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[0755] Step 6:

[0756] The user uses the interface on their device to input information about the program they want to replay (title, broadcast year, cast, etc.), and the system processes the request based on this information.

[0757] Step 7:

[0758] The device converts the request entered by the user into JSON format and sends it to the server, which then analyzes the request and searches the database for the corresponding program data.

[0759] Step 8:

[0760] The server uses a generative AI model to fill in the gaps and generate the complete video data, generating consecutive scenes in a consistent manner while maintaining audio and video synchronization.

[0761] Step 9:

[0762] The server encodes the generated video into a viewable format (e.g., MP4 format) and stores it in a database. The stored video can then be used for streaming or download.

[0763] Step 10:

[0764] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state, and the analysis results are sent to the server when the video is generated.

[0765] Step 11:

[0766] The server receives data from the emotion engine and adjusts the parameters of the generative AI model based on the user's emotional state, generating customized video according to the user's emotions.

[0767] Step 12:

[0768] The user watches the generated video through a device that provides streaming and download options to allow the user to watch the video in the appropriate way.

[0769] Step 13:

[0770] As users watch videos, the emotion engine analyzes their reactions in real time and transmits the information to the server, which can then dynamically adjust the content and tone of the video to provide the best possible viewing experience based on the user's emotions.

[0771] Example 2

[0772] 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."

[0773] In systems that reproduce past TV programs, there is a need for a method that not only plays back existing data but also provides a customized experience based on the user's emotions. Another challenge is to efficiently perform preprocessing such as noise removal from collected data and text conversion of audio data to achieve high-precision video generation.

[0774] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, and a means for user viewing. This makes it possible to reproduce past TV programs with high accuracy and dynamically change the video content and tone according to the user's emotions.

[0775] "Means of collecting data" refers to the means of obtaining data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[0776] "Preprocessing means" refers to the means for analyzing collected data, removing noise, deleting unnecessary parts, and preparing the data in a format suitable for a generative AI model.

[0777] "Means for training a generative AI model" refers to a means for training an AI model using preprocessed data and building a highly accurate model that can fill in missing parts.

[0778] The "means for user input" refers to a means by which a user inputs information about a program that the user wants to reproduce through a terminal and transmits the information to the server.

[0779] "Means for generating video" refers to a means for using a generative AI model to create video of a program requested by a user and to fill in missing parts to generate continuous video.

[0780] The "means for recognizing the user's emotions" is a means for analyzing the user's facial expressions, tone of voice, input text, etc., to determine the user's emotional state.

[0781] "Means for user viewing" refers to a means by which a user views the generated video through a terminal, and the emotion engine analyzes the user's reactions in real time while viewing and dynamically changes the video content.

[0782] This invention is a system that uses a generative AI model to recreate past TV programs and combines it with an emotion engine that recognizes the user's emotions. This system consists of three main elements: a server, a terminal, and a user.

[0783] The server is responsible for data collection, data preprocessing, training of the generative AI model, video generation, and execution of the emotion engine. Specifically, the server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, archive facilities, etc. Data collection uses Python libraries such as Requests and BeautifulSoup. The collected data is preprocessed to remove noise and unnecessary parts, and the audio data is converted to text using the Google Speech-to-Text API. This preprocessing prepares the data in a format suitable for the generative AI model.

[0784] The server then uses machine learning frameworks such as TensorFlow or PyTorch to train a generative AI model using the preprocessed data, which is then used to fill in gaps and generate a continuous video based on the user's request.

[0785] The user inputs information about the program they want to recreate (title, broadcast year, cast, etc.) through the device interface. This information is converted into JSON format and sent to the server. The server receives the user's request and searches for the corresponding program data in its database. It generates video using a generative AI model and converts it into a viewable format (e.g., MP4) using FFmpeg.

[0786] Additionally, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. The emotion engine can use OpenCV or TensorFlow, and this information is used to adjust the parameters of the generative AI model and taken into account when generating the video.

[0787] Users watch the generated video through their device. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. Users also have the option to stream or download the video.

[0788] As a concrete example, when a user requests a recreation of a 1970s variety show, an example prompt might be, "Please recreate the 1970s variety show 'XX Show.' The main character should be XX, and the atmosphere should be as joyful as possible." If the emotion engine determines that the user's emotional state is "joy," the generative AI model applies a filter that emphasizes happy scenes and bright tones. If the user is in a "sad" emotional state, the content of the video and the tone of the music are adjusted to match that emotion. In this way, a customized video experience is provided according to the user's emotions.

[0789] To implement the invention, specific hardware and software must be selected and appropriate parameters must be adjusted, but highly accurate reproduction and customization can be achieved by applying the processing steps and tools described above.

[0790] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0791] Step 1:

[0792] The server collects the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archives. Specifically, the server uses Python Requests and BeautifulSoup to perform web scraping and obtain data from APIs. The server receives a URL or API endpoint as input and obtains the collected data as output.

[0793] Step 2:

[0794] The server preprocesses the collected data. First, it converts the audio data into text using the Google Speech-to-Text API. Next, it compares the resulting text with a script to identify missing parts. It also uses regular expressions to remove noise and unnecessary characters. The collected data is given as input, and preprocessed data is obtained as output. Specifically, it performs processes such as noise removal, removal of unnecessary parts, and conversion of audio data into text.

[0795] Step 3:

[0796] The server uses the preprocessed data to train a generative AI model. It uses TensorFlow or PyTorch to build the model and sets hyperparameters (e.g., learning rate, number of epochs) to train it. The preprocessed data is given as input, and a trained generative AI model is obtained as output. Specifically, data loading, model definition, training, and validation are performed.

[0797] Step 4:

[0798] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The input information is converted to JSON format and sent to the server. The user's program information is given as input, and JSON format data is obtained as output. Specifically, form input, JSON conversion, and data transmission to the server are performed.

[0799] Step 5:

[0800] The server receives a user request and searches for the corresponding program data in a database. It uses a generative AI model to fill in missing parts and generate a continuous video. The generated video is converted into a viewable format (e.g., MP4) using FFmpeg. The user request is given as input, and the generated video file is obtained as output. Specifically, a database query, video generation, and encoding are performed.

[0801] Step 6:

[0802] The emotion engine built into the server analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. It receives real-time video feeds, audio data, and text data as input, and outputs the user's emotional state. Specifically, it uses OpenCV and TensorFlow for facial recognition, voice analysis, and text analysis.

[0803] Step 7:

[0804] The user watches the generated video through their device. The emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. The generated video and the user's reactions are given as input, and a customized viewing experience is output. Specifically, streaming playback, emotion analysis, and video content adjustment are performed.

[0805] (Application example 2)

[0806] 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."

[0807] When recreating past TV programs, it is difficult to customize the visual experience to match the viewer's emotions. Furthermore, to dynamically adjust the visual experience, it is necessary to recognize the user's emotions in real time and adjust the output of the generative AI model accordingly. Such a system does not currently exist, making it difficult to optimize the user's visual experience based on their individual emotional state.

[0808] The specific processing by the specific 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 a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, a means for dynamically adjusting the visual experience, and a means for user viewing. This makes it possible to customize the visual experience according to the user's emotions.

[0809] "Means of collecting data" refers to devices or systems for obtaining necessary data such as video, audio, scripts, etc. from the Internet, digital archives, etc.

[0810] "Preprocessing means" refers to devices and algorithms that analyze and filter collected data, remove noise and unnecessary parts, and provide high-quality data to the generative AI model.

[0811] A "means for training a generative AI model" is a device or algorithm that trains an artificial intelligence using preprocessed data, generating a model and adjusting hyperparameters.

[0812] The "means for user input" refers to an interface or device that allows the user to input information about the program that the user wants to reproduce, and that transmits the input information to the server.

[0813] "Means for generating video" refers to devices or algorithms that use generative AI models to fill in missing parts and create continuous video based on a user's request.

[0814] A "means for recognizing user emotions" is a device or algorithm that analyzes a user's facial expressions, tone of voice, and input text to determine the user's emotional state.

[0815] A "means for dynamically adjusting the visual experience" is a device or algorithm that adjusts the content and tone of a video in real time based on the user's emotions to provide the user with an optimal visual experience.

[0816] "Means for user viewing" refers to a terminal or system that allows a user to view the generated video, and that provides streaming playback or download options.

[0817] This invention is a system that reproduces past television programs and dynamically customizes the visual experience by recognizing the user's emotions. This system includes means for collecting various data, preprocessing it, and training a generative AI model, means for the user to provide input information, means for generating images, means for recognizing the user's emotions, means for dynamically adjusting the visual experience, and means for the user to view.

[0818] The server first collects the necessary data, such as video, audio, and scripts, from the internet or digital archives using APIs and scraping techniques. The collected data is then analyzed and filtered in a pre-processing stage to remove noise and unnecessary parts.

[0819] Next, the data is prepared for training a generative AI model using frameworks such as TensorFlow or PyTorch, including tuning hyperparameters. The preprocessed data is used to train the model, which then builds a generative AI model that fills in the gaps.

[0820] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). This information is converted to JSON format on the device and sent to the server. The server searches the database based on the information received from the user and retrieves the corresponding program data.

[0821] During the video generation stage, the server uses a generative AI model to fill in missing parts and create a continuous video. The generated video is then encoded and converted into a viewable format (e.g., MP4). At this time, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine their emotional state. Emotional information is fed back to the generative AI model, which dynamically adjusts the video generation parameters.

[0822] Users view the generated video using a viewing device such as a smartphone or head-mounted display. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be adjusted on the fly. The device offers streaming playback and download options to enhance the viewing experience.

[0823] As a concrete example, consider a user requesting a re-enactment of the 1970s variety show "Owarai Club." If the user smiles while watching, the emotion engine will recognize this and add fun moments and light-hearted jokes to the generated video, providing the user with a more enjoyable visual experience.

[0824] Example prompt sentence:

[0825] "Generate footage from the 1985 variety show 'Owarai Club,' and add brighter tones and fun scenes when the user is happy."

[0826] As described above, this system combines an emotion engine with a generative AI model to provide a customized visual experience that reflects the user's emotions.

[0827] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0828] Step 1: Data collection

[0829] The server collects the necessary data, such as video, audio, and scripts, from the internet or digital archives. This is done using API or scraping technology. Specifically, the server sends an API request and receives the response data in JSON format. The received data is temporarily stored in a database. A URL or API endpoint is used as input, and the collected video, audio, and script data are obtained as output.

[0830] Step 2: Data Preprocessing

[0831] The server pre-processes the collected data. This involves parsing and filtering the data to remove noise and unwanted parts. For example, it denoises the audio data and identifies missing parts of the script. The input is the collected raw data, and the output is a high-quality parsed and filtered dataset.

[0832] Step 3: Training the generative AI model

[0833] The server uses the preprocessed data to train a generative AI model. This is done using frameworks such as TensorFlow and PyTorch, and also adjusts hyperparameters. Specifically, the dataset is input into the model as training data, and parameters are optimized to fill in missing parts. The input is the preprocessed dataset, and the output is a trained generative AI model.

[0834] Step 4: Getting User Input

[0835] The terminal provides an interface for the user to input information about the program they want to replay (title, broadcast year, cast, etc.). The user inputs the information through the terminal, and the data is converted into JSON format and sent to the server. The input is the user's replay request information, and the output is the request data in JSON format.

[0836] Step 5: Image generation

[0837] The server receives a user request and searches the database for the corresponding program data. It then uses a generative AI model to fill in the missing parts and generate a continuous video. The generated video is encoded and converted into a viewable format (e.g., MP4). The input is the user request in JSON format and the trained generative AI model, and the output is the generated video data.

[0838] Step 6: Emotion Recognition

[0839] The emotion engine analyzes the user's facial expressions, voice tone, and input text to recognize the user's emotional state. This is done by collecting data in real time using the device's camera and microphone while the user is watching and applying emotion analysis algorithms. The input is the user's real-time facial and voice data, and the output is an evaluation of the user's emotional state.

[0840] Step 7: Dynamically adjust the visual experience

[0841] The server dynamically adjusts the parameters of the generative AI model based on the emotional state data obtained from the emotion engine. Specifically, if the user is happy, adjustments are made such as adding a brighter tone or a more enjoyable scene. The input is the emotional state data and the generative AI model, and the output is the adjusted video data.

[0842] Step 8: User Viewing

[0843] The user watches the generated video through the device, which offers streaming playback and download options, allowing the user to enjoy a real-time visual experience. The input is the processed video data, and the output is the user's visual experience.

[0844] In this way, the server, the terminal, and the user can work together to provide a customized visual experience according to the user's emotions.

[0845] 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.

[0846] 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.

[0847] 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.

[0848] [Fourth embodiment]

[0849] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[0850] 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.

[0851] 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).

[0852] 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.

[0853] 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.

[0854] 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).

[0855] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0856] 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.

[0857] 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.

[0858] 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.

[0859] 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.

[0860] 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.

[0861] 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."

[0862] This invention is a system for recreating and archiving past television programs using generative AI. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing.

[0863] Specific program processing

[0864] 1. Data Collection

[0865] The server collects the necessary data such as video, audio, scripts, etc. from the Internet, libraries, and archive facilities.

[0866] The server obtains data using APIs or scraping techniques and temporarily stores it.

[0867] 2. Data Preprocessing

[0868] The server analyzes the collected data and removes noise and unnecessary parts.

[0869] The server converts the audio data into text and matches it with the script to identify missing parts.

[0870] 3. Generative Model Training

[0871] The server uses the collected data to train a generative AI model, specifically using frameworks like TensorFlow and PyTorch.

[0872] The server evaluates the model's performance every epoch during the training process and makes readjustments as needed.

[0873] 4. User Input

[0874] The user inputs information about the program they want to reproduce (program name, broadcast year, cast, etc.) through the terminal interface.

[0875] The terminal converts the user's input into JSON format and sends it to the server.

[0876] 5. Image Generation

[0877] Based on a request from a user, the server searches the database for the corresponding program data.

[0878] The server uses a generative AI model to generate a video that fills in the missing parts.

[0879] The server encodes the generated video, converts it into a viewable format, and stores it in a database.

[0880] 6. User Viewing

[0881] The user watches the generated video through a device that offers streaming playback and download options.

[0882] When a user watches a video, the device uses a cache function to provide efficient access.

[0883] Specific examples

[0884] For example, imagine a user wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. The server then collects the relevant data and fills in the missing parts with a generative AI model. The resulting video can then be streamed or downloaded and viewed by the user. This process brings a precious past show back to life in a new form, making it available for both current and future audiences to enjoy.

[0885] In this way, the present invention realizes a system that can recreate past television programs and provide them in a viewable format. This system fully utilizes the capabilities of generative AI to fill in missing parts, maintaining the continuity and quality of the video, allowing viewers to access valuable video assets from the past.

[0886] The processing flow will be explained below.

[0887] Step 1:

[0888] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to automate the retrieval of information and efficiently extract the relevant content.

[0889] Step 2:

[0890] The server analyzes and pre-processes the collected data, specifically filtering to remove noise from the video data and clean up the audio data, and uses speech recognition technology to convert the audio data into text.

[0891] Step 3:

[0892] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[0893] Step 4:

[0894] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[0895] Step 5:

[0896] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[0897] Step 6:

[0898] The user uses the interface on their terminal to input information about the program they want to reproduce (title, broadcast year, cast, etc.), which sends a specific request tailored to the user's needs to the system.

[0899] Step 7:

[0900] The device converts the request entered by the user into JSON format and sends it to the server. The server analyzes the received request and searches the database for the corresponding program data.

[0901] Step 8:

[0902] The server uses a generative AI model to fill in the gaps and generate the complete video data, using technology to generate consecutive scenes in a consistent manner while maintaining synchronization between audio and video.

[0903] Step 9:

[0904] The server encodes the generated video into a viewable format (e.g., MP4) and stores it in a database, where it can be instantly streamed or downloaded.

[0905] Step 10:

[0906] Users view the generated video through a device that offers streaming playback and download options, allowing users to freely access the video.

[0907] Step 11:

[0908] When a user watches a video, the device uses a cache function to shorten the video loading time and provide smooth playback. Any user feedback is sent to the server to help improve the system.

[0909] In this way, the server, terminal, and user play their respective roles at each step, realizing a system for reproducing past television programs.

[0910] Example 1

[0911] 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."

[0912] Many past television programs have been lost over the years, resulting in the loss of valuable cultural assets for present and future viewers. Restoring lost programs using conventional technology has been difficult and often requires advanced expertise and equipment. As a result, there is a problem in that opportunities for the general public to easily watch these programs are limited.

[0913] 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.

[0914] In this invention, the server includes a means for collecting data, a means for analyzing and filtering the data, and a means for training a generative AI model, which makes it possible to complement missing parts of past TV programs and reproduce them in a viewable form.

[0915] "Means of collecting data" refers to the means of collecting the necessary data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[0916] "Means for analyzing and filtering data" refers to means for removing noise and unnecessary parts from collected data, converting voice data into text, and matching it with a script.

[0917] "Means for training a generative AI model" means means for utilizing collected and preprocessed data to train a generative AI model and obtain highly accurate output.

[0918] "Means for users to input information" refers to an interface for users to input information about the program they wish to reproduce (e.g., program name, broadcast year, cast, etc.) and means for processing the input content.

[0919] "Means for generating images" refers to means for generating images that fill in missing parts using a pre-processed and trained generative AI model.

[0920] "Means for storing the generated video and converting it into a viewable format" refers to means for encoding the video created by the generative AI model, converting it into a common viewing format, and storing it.

[0921] The "means for user viewing" refers to an interface and system that allows a user to view the generated video by streaming playback or downloading it.

[0922] This invention is a system for recreating and archiving past television programs using generative AI. This system includes means for collecting data, means for preprocessing, means for training a generative AI model, means for users to input information, means for generating video, means for saving the generated video and converting it into a viewable format, and means for users to view it.

[0923] The server first uses a data collection method to collect the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archive facilities. Specifically, it obtains the data using APIs (e.g., YouTube Data API or open databases) or scraping technologies (e.g., Scrapy or Selenium), and temporarily stores them.

[0924] The server then uses data analysis and filtering techniques to remove noise and unwanted parts from the collected data. At this stage, FFmpeg and OpenCV are used to process the video data, and noise reduction techniques are used to process the audio data. The audio data is also converted to text using the Google Cloud Speech-to-Text API, which is then matched against a script to identify missing parts.

[0925] The server then uses a method to train the generative AI model using the preprocessed dataset. Specifically, it uses TensorFlow or PyTorch to train the model and obtain highly accurate output. During the training process, the model's performance is evaluated after each epoch, and techniques such as early stopping are applied to prevent overfitting.

[0926] The user inputs the information of the program they want to reproduce through the terminal interface using the information input means. The input content is converted to JSON format and sent to the server using the HTTPS protocol. For example, the user inputs the following prompt sentence:

[0927] "Please recreate a segment from a variety show that aired in 1975. The cast is a well-known celebrity."

[0928] The server uses a video generation method to generate a continuous video based on a user request, filling in missing parts with a generative AI model. The generated video is then encoded and converted into a viewable MP4 format. Codecs used include H.264 and HEVC. The generated video is then stored in the server's database.

[0929] Finally, the user uses a viewing means to view the generated video on a device. The device provides streaming playback or download options and utilizes a streaming server to efficiently deliver the generated video. It also uses a streaming cache function while the user is viewing to provide a smooth viewing experience.

[0930] In this way, the present invention realizes a system that can reproduce past television programs and provide them in a viewable format, thereby enabling viewers to easily access valuable video assets from the past.

[0931] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0932] Program processing steps

[0933] Step 1: Data collection

[0934] Description: The server collects the necessary data (video, audio, scripts, etc.) from the internet, libraries, and archives, using APIs (e.g., open databases) and scraping techniques (e.g., Scrapy and Selenium).

[0935] Input: Raw data from the internet, libraries and archival facilities.

[0936] Data processing / data calculation: Data acquisition via API or scraping.

[0937] Output: Captured video data, audio data, and script data.

[0938] Step 2: Data Preprocessing

[0939] Description: The server analyzes the collected data and removes noise and unnecessary parts. Specifically, it uses FFmpeg and OpenCV for video data and applies noise reduction technology to audio data. It also converts audio data to text using the Google Cloud Speech-to-Text API and matches it with the script to identify missing parts.

[0940] Input: Captured video data, audio data, and script data.

[0941] Data processing / data calculation: noise removal, speech-to-text conversion, text-to-script matching.

[0942] Output: Preprocessed video data, text data, and missing part information.

[0943] Step 3: Generative model training

[0944] Description: The server trains a generative AI model using preprocessed data. Specifically, it uses TensorFlow or PyTorch to split the dataset into training data and validation data and train the model. It evaluates the model's performance every epoch and applies techniques such as early stopping to prevent overfitting.

[0945] Input: Preprocessed video data, text data, and missing part information.

[0946] Data processing / data computation: model training, performance evaluation, and hyperparameter tuning.

[0947] Output: A trained generative AI model.

[0948] Step 4: User Input

[0949] Description: The user inputs the information of the program they want to reproduce through the device interface. The device converts the user's input into JSON format and sends it to the server using the HTTPS protocol.

[0950] Input: Program information entered by the user (e.g., program name, broadcast year, cast).

[0951] Data processing / data calculation: Convert user input into JSON format and send to the server.

[0952] Output: The request data sent to the server.

[0953] Step 5: Image generation

[0954] Description: The server searches the database for the relevant program data based on the user's request, and generates a video that fills in the missing parts using a generative AI model.The generated video is then encoded using a codec such as H.264 or HEVC, converted to MP4 format, and saved in the database.

[0955] Input: User request data, trained generative AI model, and program data from the database.

[0956] Data processing / data calculation: database search, generation of complementary video, video encoding, conversion to MP4 format.

[0957] Output: The generated video data.

[0958] Step 6: User Viewing

[0959] Description: The user uses a device to watch the generated video. The device uses a streaming server (e.g., NGINX or Wowza) to stream the video or provides a download option. During viewing, a cache function is used to provide a smooth viewing experience.

[0960] Input: Generated video data.

[0961] Data processing / data calculation: streaming distribution, cache processing.

[0962] Output: The video that the user can see.

[0963] (Application example 1)

[0964] 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."

[0965] There is a need to recreate valuable past television programs and video assets so that they can be viewed by current and future viewers. However, over the years, the original video data can deteriorate or become missing. Furthermore, when recreating past video, noise and unnecessary information are often mixed in, significantly reducing viewing quality. Therefore, a system is needed that can efficiently recreate past television programs and provide them to viewers with high quality.

[0966] 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.

[0967] In this invention, the server includes means for collecting data, means for preprocessing, means for training a generative AI model, means for user input, means for generating video, means for user viewing, means for analyzing the collected data and processing it to reproduce video of a specified program, means for providing an interface for users to request past video, and means for converting video generated using hardware into a viewable format, thereby making it possible to reproduce past TV programs in high quality and efficiently provide video requested by viewers.

[0968] "Data collection means" refers to the devices and methods used to collect the necessary data from the Internet and archive facilities.

[0969] "Preprocessing means" refers to a device or method that analyzes collected data and removes noise and unnecessary parts.

[0970] A "means for training a generative AI model" is a device or method that uses collected data to train an AI model and acquire knowledge to fill in any gaps.

[0971] "User input means" refers to a device or method for a user to input information about a program that the user wants to reproduce (such as the program name, broadcast year, and cast members).

[0972] "Image generation means" refers to a device or method that uses preprocessed data and a generative AI model to generate images that fill in missing parts.

[0973] "User viewing means" refers to a device or method that allows a user to view the generated video.

[0974] The "collected data analysis means" refers to a device or method that analyzes the collected data and performs the processing necessary to reproduce the video of the specified program.

[0975] The "user interface providing means" refers to a device or method that provides an interface for a user to request past video.

[0976] "Video conversion means" refers to a device or method that converts video generated using hardware into a viewable format.

[0977] This invention is a system that uses generative AI to recreate and archive past television programs. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, and a means for user viewing. Specific embodiments of these means are described below.

[0978] First, in the data collection process, the server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. This is achieved using APIs and scraping technology. The collected data is temporarily stored and used for subsequent processing.

[0979] Next, the pre-processing means analyzes the data collected by the server and removes noise and unnecessary parts. Specifically, it performs frame analysis of the collected video data and converts audio data into text. This converts the data into a format suitable for video complementation.

[0980] The generative AI model training method uses the collected data to train the generative AI model. Using frameworks like TensorFlow or PyTorch, the model's performance is evaluated every epoch and retuned as needed. As the model learns, it acquires knowledge to fill in missing parts of the video.

[0981] The user input means allows the user to input information about the program they want to reproduce (such as the program name, broadcast year, and cast members) through the device's interface. The device converts the user's input into JSON format and sends it to the server, allowing the server to efficiently search for specific program data.

[0982] In the video generation method, the server uses the generative AI model to generate videos that fill in missing parts. These videos are encoded into a format that can be viewed by users and stored in a database.

[0983] Finally, as a means for user viewing, users can view the generated video through their devices, with streaming playback and download options available, and a cache function for efficient access.

[0984] As a concrete example, consider a user who wants to recreate a variety show that aired in the 1970s. The user uses the device interface to input the show name and broadcast year and submits the request. In this case, the prompt sentence is "Please recreate the 1970 variety show 'Retro Variety Show'." The server collects the relevant data and fills in the missing parts with a generative AI model. As a result, the user can stream or download the generated video and watch it.

[0985] As described above, the present invention makes it possible to reproduce past television programs with high quality and efficiently provide the video requested by the viewer.

[0986] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0987] Step 1: Data collection

[0988] The server collects the necessary data, such as video, audio, and scripts, from the internet and archive facilities. Specifically, it uses APIs and scraping technology to obtain data related to the specified program. The input is the program name and broadcast year, and the output is temporarily stored video, audio, and text data.

[0989] Step 2: Data Preprocessing

[0990] The server analyzes the collected data and processes it to remove noise and unnecessary parts. Specifically, it uses OpenCV to analyze frames of video data and applies a noise removal filter. Additionally, audio data is converted into text using speech recognition technology and matched with a script. The input is the raw collected data, and the output is the clean data that has been analyzed and filtered.

[0991] Step 3: Training the generative AI model

[0992] The server uses the preprocessed data to train a generative AI model. This uses frameworks such as TensorFlow and PyTorch. Specifically, it pairs video and text data to create a training set, and then trains the model based on that. The input is the preprocessed data, and the output is a fully trained generative AI model.

[0993] Step 4: User Input

[0994] The user inputs information about the program they want to play (such as the program name, broadcast year, and cast) through the device's interface. The device converts this input into JSON format and sends it to the server. The input is the user's program request information, and the output is JSON format data.

[0995] Step 5: Image generation

[0996] Based on a user request, the server uses a generative AI model to generate a video with missing parts filled in. Specifically, it uses a trained model to generate reconstructed video frames from clean data. The input is the user request and clean data, and the output is the filled video data.

[0997] Step 6: User Viewing

[0998] The user watches the generated video through a device. The device provides a player for streaming the video, and also offers a download option. The input is the generated video data, and the output is a viewable video.

[0999] As described above, by performing appropriate data processing and data calculations based on the input data at each step, a system is realized that reproduces past television programs with high quality and provides the video requested by the user.

[1000] 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.

[1001] This invention is a system that uses generative AI to recreate past television programs and combines it with an emotion engine that recognizes user emotions. This system includes a means for collecting data, a means for preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for user viewing, and an emotion engine that recognizes user emotions.

[1002] Specific program processing

[1003] 1. Data Collection

[1004] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives using APIs and scraping technology.

[1005] 2. Data Preprocessing

[1006] The server analyzes the collected data, filtering out noise and unnecessary parts, converts the audio data to text, and matches it with a script to identify missing parts, thereby preparing the data for subsequent input into a generative AI model.

[1007] 3. Generative Model Training

[1008] The server trains a generative AI model using the preprocessed data. Using the prepared dataset, it builds a generative AI model that fills in the gaps. The training uses the TensorFlow and PyTorch frameworks, and also adjusts hyperparameters.

[1009] 4. User Input

[1010] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The device converts the input information into JSON format and sends it to the server.

[1011] 5. Image Generation

[1012] The server receives the user's request, searches the database for the corresponding program data, uses a generative AI model to fill in the gaps and create a continuous video, and then encodes the video and converts it into a viewable format (e.g., MP4).

[1013] 6. Operation of the Emotion Engine

[1014] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. This information is taken into account when generating the video, and the parameters of the generative AI model are adjusted as needed.

[1015] 7. User Viewing

[1016] Users watch the generated video through the device. An emotion engine analyzes users' reactions in real time, and the content and tone of the video may be dynamically changed. The device offers streaming and download options to enhance the viewing experience.

[1017] Specific examples

[1018] For example, consider a user requesting a re-enactment of a 1970s variety show. If the emotion engine determines that the user is in a positive emotional state (e.g., joy), the generative AI model applies a filter that emphasizes more joyful scenes and brighter tones. On the other hand, if the user is in a sad emotional state, the video content and musical tone are adjusted to match that emotion. In this way, a customized video experience is provided based on the user's emotions.

[1019] This system not only makes precious programs from the past available to current and future audiences in new formats, but also customizes the visual experience based on the user's emotional state. The collaboration between generative AI and the emotion engine makes it possible to provide even more compelling visual content.

[1020] The processing flow will be explained below.

[1021] Step 1:

[1022] The server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, and archives, using APIs and scraping technology to efficiently extract the relevant content.

[1023] Step 2:

[1024] The server analyzes the collected data and removes noise and unnecessary parts, specifically by performing noise reduction processing on video data and filtering on audio data, and also by using voice recognition technology to convert audio data into text.

[1025] Step 3:

[1026] The server matches the preprocessed audio with the script data, identifies gaps, annotates them, and prepares the data for subsequent input into a generative AI model.

[1027] Step 4:

[1028] The server builds a generative AI model and trains it using preprocessed data, using deep learning frameworks such as TensorFlow and PyTorch to appropriately set the model's hyperparameters.

[1029] Step 5:

[1030] The server evaluates the performance of the trained generative AI model by using validation data to check how accurate the generated results are and, if necessary, retraining the model.

[1031] Step 6:

[1032] The user uses the interface on their device to input information about the program they want to replay (title, broadcast year, cast, etc.), and the system processes the request based on this information.

[1033] Step 7:

[1034] The device converts the request entered by the user into JSON format and sends it to the server, which then analyzes the request and searches the database for the corresponding program data.

[1035] Step 8:

[1036] The server uses a generative AI model to fill in the gaps and generate the complete video data, generating consecutive scenes in a consistent manner while maintaining audio and video synchronization.

[1037] Step 9:

[1038] The server encodes the generated video into a viewable format (e.g., MP4 format) and stores it in a database. The stored video can then be used for streaming or download.

[1039] Step 10:

[1040] The emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state, and the analysis results are sent to the server when the video is generated.

[1041] Step 11:

[1042] The server receives data from the emotion engine and adjusts the parameters of the generative AI model based on the user's emotional state, generating customized video according to the user's emotions.

[1043] Step 12:

[1044] The user watches the generated video through a device that provides streaming and download options to allow the user to watch the video in the appropriate way.

[1045] Step 13:

[1046] As users watch videos, the emotion engine analyzes their reactions in real time and transmits the information to the server, which can then dynamically adjust the content and tone of the video to provide the best possible viewing experience based on the user's emotions.

[1047] Example 2

[1048] 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."

[1049] In systems that reproduce past TV programs, there is a need for a method that not only plays back existing data but also provides a customized experience based on the user's emotions. Another challenge is to efficiently perform preprocessing such as noise removal from collected data and text conversion of audio data to achieve high-precision video generation.

[1050] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, and a means for user viewing. This makes it possible to reproduce past TV programs with high accuracy and dynamically change the video content and tone according to the user's emotions.

[1051] "Means of collecting data" refers to the means of obtaining data such as video, audio, and scripts from the Internet, libraries, and archive facilities.

[1052] "Preprocessing means" refers to the means for analyzing collected data, removing noise, deleting unnecessary parts, and preparing the data in a format suitable for a generative AI model.

[1053] "Means for training a generative AI model" refers to a means for training an AI model using preprocessed data and building a highly accurate model that can fill in missing parts.

[1054] The "means for user input" refers to a means by which a user inputs information about a program that the user wants to reproduce through a terminal and transmits the information to the server.

[1055] "Means for generating video" refers to a means for using a generative AI model to create video of a program requested by a user and to fill in missing parts to generate continuous video.

[1056] The "means for recognizing the user's emotions" is a means for analyzing the user's facial expressions, tone of voice, input text, etc., to determine the user's emotional state.

[1057] "Means for user viewing" refers to a means by which a user views the generated video through a terminal, and the emotion engine analyzes the user's reactions in real time while viewing and dynamically changes the video content.

[1058] This invention is a system that uses a generative AI model to recreate past TV programs and combines it with an emotion engine that recognizes the user's emotions. This system consists of three main elements: a server, a terminal, and a user.

[1059] The server is responsible for data collection, data preprocessing, training of the generative AI model, video generation, and execution of the emotion engine. Specifically, the server collects the necessary data, such as video, audio, and scripts, from the internet, libraries, archive facilities, etc. Data collection uses Python libraries such as Requests and BeautifulSoup. The collected data is preprocessed to remove noise and unnecessary parts, and the audio data is converted to text using the Google Speech-to-Text API. This preprocessing prepares the data in a format suitable for the generative AI model.

[1060] The server then uses machine learning frameworks such as TensorFlow or PyTorch to train a generative AI model using the preprocessed data, which is then used to fill in gaps and generate a continuous video based on the user's request.

[1061] The user inputs information about the program they want to recreate (title, broadcast year, cast, etc.) through the device interface. This information is converted into JSON format and sent to the server. The server receives the user's request and searches for the corresponding program data in its database. It generates video using a generative AI model and converts it into a viewable format (e.g., MP4) using FFmpeg.

[1062] Additionally, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. The emotion engine can use OpenCV or TensorFlow, and this information is used to adjust the parameters of the generative AI model and taken into account when generating the video.

[1063] Users watch the generated video through their device. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. Users also have the option to stream or download the video.

[1064] As a concrete example, when a user requests a recreation of a 1970s variety show, an example prompt might be, "Please recreate the 1970s variety show 'XX Show.' The main character should be XX, and the atmosphere should be as joyful as possible." If the emotion engine determines that the user's emotional state is "joy," the generative AI model applies a filter that emphasizes happy scenes and bright tones. If the user is in a "sad" emotional state, the content of the video and the tone of the music are adjusted to match that emotion. In this way, a customized video experience is provided according to the user's emotions.

[1065] To implement the invention, specific hardware and software must be selected and appropriate parameters must be adjusted, but highly accurate reproduction and customization can be achieved by applying the processing steps and tools described above.

[1066] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1067] Step 1:

[1068] The server collects the necessary data, such as video, audio, and scripts, from the Internet, libraries, and archives. Specifically, the server uses Python Requests and BeautifulSoup to perform web scraping and obtain data from APIs. The server receives a URL or API endpoint as input and obtains the collected data as output.

[1069] Step 2:

[1070] The server preprocesses the collected data. First, it converts the audio data into text using the Google Speech-to-Text API. Next, it compares the resulting text with a script to identify missing parts. It also uses regular expressions to remove noise and unnecessary characters. The collected data is given as input, and preprocessed data is obtained as output. Specifically, it performs processes such as noise removal, removal of unnecessary parts, and conversion of audio data into text.

[1071] Step 3:

[1072] The server uses the preprocessed data to train a generative AI model. It uses TensorFlow or PyTorch to build the model and sets hyperparameters (e.g., learning rate, number of epochs) to train it. The preprocessed data is given as input, and a trained generative AI model is obtained as output. Specifically, data loading, model definition, training, and validation are performed.

[1073] Step 4:

[1074] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). The input information is converted to JSON format and sent to the server. The user's program information is given as input, and JSON format data is obtained as output. Specifically, form input, JSON conversion, and data transmission to the server are performed.

[1075] Step 5:

[1076] The server receives a user request and searches for the corresponding program data in a database. It uses a generative AI model to fill in missing parts and generate a continuous video. The generated video is converted into a viewable format (e.g., MP4) using FFmpeg. The user request is given as input, and the generated video file is obtained as output. Specifically, a database query, video generation, and encoding are performed.

[1077] Step 6:

[1078] The emotion engine built into the server analyzes the user's facial expressions, voice tone, and input text to determine the user's emotional state. It receives real-time video feeds, audio data, and text data as input, and outputs the user's emotional state. Specifically, it uses OpenCV and TensorFlow for facial recognition, voice analysis, and text analysis.

[1079] Step 7:

[1080] The user watches the generated video through their device. The emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be dynamically changed. The generated video and the user's reactions are given as input, and a customized viewing experience is output. Specifically, streaming playback, emotion analysis, and video content adjustment are performed.

[1081] (Application example 2)

[1082] 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."

[1083] When recreating past TV programs, it is difficult to customize the visual experience to match the viewer's emotions. Furthermore, to dynamically adjust the visual experience, it is necessary to recognize the user's emotions in real time and adjust the output of the generative AI model accordingly. Such a system does not currently exist, making it difficult to optimize the user's visual experience based on their individual emotional state.

[1084] The specific processing by the specific 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 a means for collecting data, a means for performing preprocessing, a means for training a generative AI model, a means for user input, a means for generating video, a means for recognizing user emotions, a means for dynamically adjusting the visual experience, and a means for user viewing. This makes it possible to customize the visual experience according to the user's emotions.

[1085] "Means of collecting data" refers to devices or systems for obtaining necessary data such as video, audio, scripts, etc. from the Internet, digital archives, etc.

[1086] "Preprocessing means" refers to devices and algorithms that analyze and filter collected data, remove noise and unnecessary parts, and provide high-quality data to the generative AI model.

[1087] A "means for training a generative AI model" is a device or algorithm that trains an artificial intelligence using preprocessed data, generating a model and adjusting hyperparameters.

[1088] The "means for user input" refers to an interface or device that allows the user to input information about the program that the user wants to reproduce, and that transmits the input information to the server.

[1089] "Means for generating video" refers to devices or algorithms that use generative AI models to fill in missing parts and create continuous video based on a user's request.

[1090] A "means for recognizing user emotions" is a device or algorithm that analyzes a user's facial expressions, tone of voice, and input text to determine the user's emotional state.

[1091] A "means for dynamically adjusting the visual experience" is a device or algorithm that adjusts the content and tone of a video in real time based on the user's emotions to provide the user with an optimal visual experience.

[1092] "Means for user viewing" refers to a terminal or system that allows a user to view the generated video, and that provides streaming playback or download options.

[1093] This invention is a system that reproduces past television programs and dynamically customizes the visual experience by recognizing the user's emotions. This system includes means for collecting various data, preprocessing it, and training a generative AI model, means for the user to provide input information, means for generating images, means for recognizing the user's emotions, means for dynamically adjusting the visual experience, and means for the user to view.

[1094] The server first collects the necessary data, such as video, audio, and scripts, from the internet or digital archives using APIs and scraping techniques. The collected data is then analyzed and filtered in a pre-processing stage to remove noise and unnecessary parts.

[1095] Next, the data is prepared for training a generative AI model using frameworks such as TensorFlow or PyTorch, including tuning hyperparameters. The preprocessed data is used to train the model, which then builds a generative AI model that fills in the gaps.

[1096] The user uses the device interface to input information about the program they want to reproduce (title, broadcast year, cast, etc.). This information is converted to JSON format on the device and sent to the server. The server searches the database based on the information received from the user and retrieves the corresponding program data.

[1097] During the video generation stage, the server uses a generative AI model to fill in missing parts and create a continuous video. The generated video is then encoded and converted into a viewable format (e.g., MP4). At this time, an emotion engine analyzes the user's facial expressions, voice tone, and input text to determine their emotional state. Emotional information is fed back to the generative AI model, which dynamically adjusts the video generation parameters.

[1098] Users view the generated video using a viewing device such as a smartphone or head-mounted display. An emotion engine analyzes the user's reactions in real time, and the content and tone of the video may be adjusted on the fly. The device offers streaming playback and download options to enhance the viewing experience.

[1099] As a concrete example, consider a user requesting a re-enactment of the 1970s variety show "Owarai Club." If the user smiles while watching, the emotion engine will recognize this and add fun moments and light-hearted jokes to the generated video, providing the user with a more enjoyable visual experience.

[1100] Example prompt sentence:

[1101] "Generate footage from the 1985 variety show 'Owarai Club,' and add brighter tones and fun scenes when the user is happy."

[1102] As described above, this system combines an emotion engine with a generative AI model to provide a customized visual experience that reflects the user's emotions.

[1103] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1104] Step 1: Data collection

[1105] The server collects the necessary data, such as video, audio, and scripts, from the internet or digital archives. This is done using API or scraping technology. Specifically, the server sends an API request and receives the response data in JSON format. The received data is temporarily stored in a database. A URL or API endpoint is used as input, and the collected video, audio, and script data are obtained as output.

[1106] Step 2: Data Preprocessing

[1107] The server pre-processes the collected data. This involves parsing and filtering the data to remove noise and unwanted parts. For example, it denoises the audio data and identifies missing parts of the script. The input is the collected raw data, and the output is a high-quality parsed and filtered dataset.

[1108] Step 3: Training the generative AI model

[1109] The server uses the preprocessed data to train a generative AI model. This is done using frameworks such as TensorFlow and PyTorch, and also adjusts hyperparameters. Specifically, the dataset is input into the model as training data, and parameters are optimized to fill in missing parts. The input is the preprocessed dataset, and the output is a trained generative AI model.

[1110] Step 4: Getting User Input

[1111] The terminal provides an interface for the user to input information about the program they want to replay (title, broadcast year, cast, etc.). The user inputs the information through the terminal, and the data is converted into JSON format and sent to the server. The input is the user's replay request information, and the output is the request data in JSON format.

[1112] Step 5: Image generation

[1113] The server receives a user request and searches the database for the corresponding program data. It then uses a generative AI model to fill in the missing parts and generate a continuous video. The generated video is encoded and converted into a viewable format (e.g., MP4). The input is the user request in JSON format and the trained generative AI model, and the output is the generated video data.

[1114] Step 6: Emotion Recognition

[1115] The emotion engine analyzes the user's facial expressions, voice tone, and input text to recognize the user's emotional state. This is done by collecting data in real time using the device's camera and microphone while the user is watching and applying emotion analysis algorithms. The input is the user's real-time facial and voice data, and the output is an evaluation of the user's emotional state.

[1116] Step 7: Dynamically adjust the visual experience

[1117] The server dynamically adjusts the parameters of the generative AI model based on the emotional state data obtained from the emotion engine. Specifically, if the user is happy, adjustments are made such as adding a brighter tone or a more enjoyable scene. The input is the emotional state data and the generative AI model, and the output is the adjusted video data.

[1118] Step 8: User Viewing

[1119] The user watches the generated video through the device, which offers streaming playback and download options, allowing the user to enjoy a real-time visual experience. The input is the processed video data, and the output is the user's visual experience.

[1120] In this way, the server, the terminal, and the user can work together to provide a customized visual experience according to the user's emotions.

[1121] 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.

[1122] 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.

[1123] 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.

[1124] 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.

[1125] 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.

[1126] 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.

[1127] 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).

[1128] 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.

[1129] 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."

[1130] 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.

[1131] 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).

[1132] 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.

[1133] 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.

[1134] 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.

[1135] 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.

[1136] 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.

[1137] 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.

[1138] 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.

[1139] 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.

[1140] 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.

[1141] 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.

[1142] The following is further disclosed regarding the above embodiment.

[1143] (Claim 1)

[1144] a means of collecting data;

[1145] A means for performing pre-processing;

[1146] A means of training the generative AI model;

[1147] a means for user input;

[1148] a means for generating an image;

[1149] a means for a user to view;

[1150] A system including:

[1151] (Claim 2)

[1152] 2. The system of claim 1, wherein the preprocessing means analyzes and filters the data to remove noise and unwanted portions.

[1153] (Claim 3)

[1154] The system of claim 1, wherein the generative AI model has the ability to convert voice data into text and match it with a script.

[1155] "Example 1"

[1156] (Claim 1)

[1157] a means of collecting data;

[1158] means for analyzing and filtering the data;

[1159] A means of training the generative AI model;

[1160] a means for a user to input information;

[1161] a means for generating an image;

[1162] means for storing the generated video and converting it into a viewable format;

[1163] a means for a user to view;

[1164] A system including:

[1165] (Claim 2)

[1166] 2. The system of claim 1, wherein the preprocessing means analyzes and filters the data to remove noise and unwanted portions.

[1167] (Claim 3)

[1168] The system described in claim 1, characterized in that the generative AI model has the function of converting voice data into text, matching it with a script, and completing missing parts.

[1169] "Application Example 1"

[1170] (Claim 1)

[1171] a means of collecting data;

[1172] A means for performing pre-processing;

[1173] A means of training the generative AI model;

[1174] a means for user input;

[1175] a means for generating an image;

[1176] a means for a user to view;

[1177] A means for analyzing the collected data and processing it to reproduce the video of the designated program;

[1178] means for providing an interface for a user to request past footage;

[1179] means for converting the hardware-generated video into a viewable format;

[1180] A system including:

[1181] (Claim 2)

[1182] 2. The system according to claim 1, wherein the pre-processing means analyzes and filters the data, removes noise and unnecessary parts, and performs processing necessary to complement the image.

[1183] (Claim 3)

[1184] The system described in claim 1, characterized in that the generative AI model has the function of converting voice data into text, matching it with a script, and completing any missing parts.

[1185] "Example 2: Combining Emotion Engines"

[1186] (Claim 1)

[1187] a means of collecting data;

[1188] A means for performing pre-processing;

[1189] A means of training the generative AI model;

[1190] a means for user input;

[1191] a means for generating an image;

[1192] means for recognizing a user's emotion;

[1193] a means for a user to view;

[1194] A system including:

[1195] (Claim 2)

[1196] 2. The system of claim 1, wherein the preprocessing means analyzes and filters the data to remove noise and unwanted portions.

[1197] (Claim 3)

[1198] The system of claim 1, wherein the generative AI model has the ability to convert voice data into text and match it with a script.

[1199] (Claim 4)

[1200] 10. The system of claim 1, wherein the means for recognizing a user's emotion analyzes the user's facial expression, tone of voice, and input text.

[1201] (Claim 5)

[1202] 5. The system of claim 4, wherein the means for recognizing user emotions adjusts parameters of the generative AI model based on the user's emotional state.

[1203] "Application example 2 when combining emotion engines"

[1204] New Claims

[1205] (Claim 1)

[1206] a means of collecting data;

[1207] A means for performing pre-processing;

[1208] A means of training the generative AI model;

[1209] a means for user input;

[1210] a means for generating an image;

[1211] means for recognizing a user's emotion;

[1212] a means for dynamically adjusting the visual experience;

[1213] a means for a user to view;

[1214] A system including:

[1215] (Claim 2)

[1216] 2. The system of claim 1, wherein the preprocessing means analyzes and filters the data to remove noise and unwanted portions.

[1217] (Claim 3)

[1218] The system of claim 1, characterized in that the generative AI model has the ability to convert voice data into text, match it with a script, and dynamically adjust the visual experience according to the user's emotions. [Explanation of symbols]

[1219] 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 of collecting data; A means for performing pre-processing; A means of training the generative AI model; a means for user input; a means for generating an image; a means for a user to view; A system including:

2. 2. The system of claim 1, wherein the preprocessing means analyzes and filters the data to remove noise and unwanted portions.

3. 2. The system of claim 1, wherein the generative AI model has the function of converting voice data into text and matching it with a script.

Images