system

Abstract

We provide the system. [Solution] A means of collecting image data from a terminal and sending it to a server, A means for analyzing image data on a server and generating captions, A means of sending the generated caption to a terminal and converting it into speech, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] As an issue faced by visually impaired persons in understanding image content on the Internet or social network services, when alternative text is not set, it is difficult to grasp the content of an image. This issue leads to restrictions on information access, causing the problem that users with visual impairments cannot equally utilize information on the Internet.

Means for Solving the Problems

[0005] This invention enables efficient analysis of images on the web by providing a means for collecting image data from a terminal and transmitting it to a server. The server uses an AI model to analyze the image data and generates captions based on its content. The generated captions are then transmitted to the terminal, and the content of the images can be conveyed to visually impaired individuals verbally using a means for converting the captions to speech on the terminal. This entire process allows visually impaired individuals to understand image content without alternative text in real time, thereby reducing barriers to information access.

[0006] A "terminal" is a device used by a user, and is a device that collects and transmits image data.

[0007] "Image data" refers to data that includes visual information displayed on web pages and social networking services.

[0008] A "server" is a central system that receives image data transmitted from terminals, performs analysis, and generates captions.

[0009] An "AI model" is an artificial intelligence algorithm and structure that analyzes image data and understands its content.

[0010] A "caption" is a natural language description generated to describe the content of image data.

[0011] "Voice conversion" is the process of converting generated captions into speech and communicating them to the user aurally. [Brief explanation of the drawing]

[0012] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] It is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] It is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] It is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] It is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] It is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] It is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] It shows an emotion map to which a plurality of emotions are mapped. [Figure 10] It shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Embodiments for Carrying Out the Invention

[0013] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0014] First, the language used in the following description will be explained.

[0015] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0016] In the following embodiments, the numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0017] In the following embodiments, the numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, and the like.

[0018] In the following embodiments, the numbered communication I / F (Interface) is an interface including a communication processor and 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), or Bluetooth (registered trademark), and the like.

[0019] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0020] [First Embodiment]

[0021] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0022] As shown in Figure 1, the 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.

[0023] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0024] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0025] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and 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.

[0026] 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0027] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0028] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0029] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0030] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0031] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0032] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0033] This invention provides a system that enables visually impaired individuals to understand image content on the internet and social networking services. The system includes a terminal, a server, and software with caption generation and speech conversion capabilities.

[0034] Users typically use their devices to navigate websites and social media, opening pages of interest. During this process, the device has a function to automatically detect images on the page. The detected image data is then transmitted to a server via the internet.

[0035] The server uses an AI model to analyze the received image data. This AI model has the ability to analyze the image content in detail using technologies such as object recognition and scene detection. Once the analysis is complete, the server generates text captions corresponding to the images.

[0036] The generated captions are sent back to the device and converted into audio. This conversion uses natural-sounding speech synthesis technology, allowing users to hear the content of the image in audio form.

[0037] For example, if a user opens a travel blog page and tries to view a photograph of a mountain landscape, the image is sent to the server, where a caption such as "The Alps and lush green valleys stretching out under a blue sky" is generated. The device then provides this caption to the user via audio, allowing them to understand the image's content without relying on visual information.

[0038] This system provides a barrier-free environment for visually impaired individuals, enabling them to access images without alternative text, and significantly improving their online information access.

[0039] The following describes the processing flow.

[0040] Step 1:

[0041] A user uses their device to access websites and social networking services and opens pages of interest. During this process, the user performs typical browsing actions.

[0042] Step 2:

[0043] The device automatically detects images on the opened page and temporarily retrieves the image data. The images are stored in URL or base64 format.

[0044] Step 3:

[0045] The terminal prepares to send the acquired image data to the server and sends the image data to the server via an internet connection in the form of an API request.

[0046] Step 4:

[0047] The server receives image data sent from the terminal and begins analyzing the image using an AI model. This analysis includes object recognition and scene detection.

[0048] Step 5:

[0049] The server generates image captions based on the analysis results. The generated captions describe the details of the objects and scenes in the image using natural language.

[0050] Step 6:

[0051] The server sends the generated caption back to the terminal. At this time, the caption is sent in a standard data format.

[0052] Step 7:

[0053] The terminal receives captions from the server and uses a text-to-speech engine to convert the text captions into speech.

[0054] Step 8:

[0055] The device provides the user with an audio caption via a speaker or headphones, allowing the user to understand the image content audibly.

[0056] (Example 1)

[0057] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0058] Individuals with visual impairments face difficulties in effectively understanding visual content on the internet and digital media. Furthermore, systems providing audio information as an alternative to visual information are not sufficiently developed, resulting in a lack of means to accurately convey the meaning of visual content. This situation leads to the challenge of limited information access for individuals with visual impairments.

[0059] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0060] In this invention, the server includes means for analyzing visual information and generating natural language descriptions, means for a mechanism for analyzing visual signals, and means for using sound generation technology to generate naturally sounding synthesized speech. This enables individuals with visual impairments to accurately understand visual information as sound and to smoothly access information on the internet and digital media.

[0061] A "terminal" is a computer device used to detect visual information and transmit it to an information processing device.

[0062] "Visual information" refers to visual data detected from a device as digital images or videos.

[0063] An "information processing device" is a computer device that analyzes visual information received from a terminal and generates natural language descriptions.

[0064] "Natural language description" refers to a form of text that is understandable to humans and is generated based on visual information.

[0065] "Visual signals" refer to image and video data that a device detects and analyzes.

[0066] "Sound generation technology" is a technology that converts text data into speech that sounds like human speech.

[0067] A "learning model" is a computational model that analyzes visual information based on AI technology to perform feature recognition and scene detection.

[0068] This invention provides a system for individuals with visual impairments to understand visual information on the internet and digital media as audio. The system includes a terminal, an information processing device, and software employing sound generation technology to convert visual information into audio.

[0069] Users access web pages and social media content containing visual information through their devices. These devices have the capability to detect visual information on the page and transmit this information to an information processing device. The devices used are commonly available personal computers and smart devices.

[0070] The server, as an information processing unit, is equipped with a learning model for analyzing received visual information. This model is implemented using widely used AI frameworks such as TENSORFLOW® and PyTorch, and recognizes specific features from visual information to generate natural language descriptions. For example, it performs object recognition and scene detection, and then expresses the visual information as text based on the results.

[0071] The generated natural language description is sent back to the device. The device uses sound generation technology, such as Google® Text-to-Speech API, to convert the natural language description into synthesized speech. This allows the user to understand visual information through speech.

[0072] As a concrete example, consider a case where a user views a blog post containing a photograph of a tourist destination. The device sends this photograph to the server, which generates a natural language description from the visual information, such as "a historical castle and its surrounding landscape against a blue sky." The device then provides this description to the user as synthesized speech. An example of a prompt could be the instruction, "Please create a caption that describes the content of this image in detail."

[0073] By implementing this invention, individuals with visual impairments can receive visual information as audio, significantly improving their access to information provided online.

[0074] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0075] Step 1:

[0076] The user opens internet or social media pages on their device. The device automatically detects images contained in the page. Specifically, it analyzes the HTML and CSS, This tool extracts data set as background images using tags and CSS. The input is the HTML data of a webpage, and the output is the image URL or binary data.

[0077] Step 2:

[0078] The device sends the detected image data to the server. Specifically, it sends the image data using an HTTP POST request. In this process, the image metadata may also be sent. The input is image data (e.g., JPEG or PNG format), and the output is an HTTP request that is sent to the server.

[0079] Step 3:

[0080] The server inputs the received image data into an AI model. The AI ​​model is, for example, a deep learning model using TensorFlow, and performs object recognition and scene detection. The input is image data, and the output is intermediate data describing the image's features. Specifically, a convolutional neural network (CNN) extracts the main features of the image.

[0081] Step 4:

[0082] The server generates natural language descriptions based on the output of the AI ​​model. The generative AI model analyzes feature data and generates corresponding text. For example, it might create a caption such as "A historical castle spread out under a blue sky." The input is the feature data of the AI ​​model, and the output is a natural language description.

[0083] Step 5:

[0084] The server sends the generated natural language description to the terminal. It returns an HTTP response, transmitting the natural language description to the terminal. The input is the generated natural language description, and the output is the text data sent to the terminal.

[0085] Step 6:

[0086] The device inputs the received natural language description into a speech synthesis engine and generates speech. For example, it uses the Google Text-to-Speech API to convert text into speech. The input is natural language text, and the output is synthesized speech data. Specifically, it converts the text into phonemes and generates a sound waveform.

[0087] Step 7:

[0088] The user listens to audio provided by the device. This allows the user to perceive visual information audibly. The device plays the audio through speakers or headphones. This process includes playback controls and volume adjustments. The input is synthesized audio data, and the output is the audio played back to the user.

[0089] (Application Example 1)

[0090] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0091] Visually impaired individuals face the challenge of understanding online video and image content in real time. Furthermore, general visual caption generation systems are specialized for static images only and are insufficient to assist in understanding dynamic content. Specifically, there is a need for methods to grasp the actions of characters and details of scenes in video works such as movies and television programs in real time, without relying on sight.

[0092] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0093] In this invention, the server includes means for collecting visual information from a terminal and transmitting it to a central processing unit, means for the central processing unit to analyze the visual information and generate a text description, and means for transmitting the generated text description to the terminal and outputting it as audio. This makes it possible for visually impaired people to understand visual information in real time and to enrich their experience of dynamic content.

[0094] A "terminal" is a device that collects visual information and transmits it to a central processing unit.

[0095] "Visual information" refers to visual content such as images and videos, which are analyzed by a central processing unit.

[0096] A "central processing unit" is a system that analyzes visual information transmitted from terminals and generates textual descriptions.

[0097] A "text description" is a written representation of the content of the analyzed visual information, generated to help visually impaired individuals understand its content.

[0098] A "generative model" refers to artificial intelligence technology used to generate appropriate text descriptions from visual information.

[0099] "Audio output" refers to the process of providing the generated text explanation to the user as audio.

[0100] "Real-time" refers to a state where events are processed simultaneously, meaning that information is provided without delay.

[0101] "Dynamic content" refers to content that includes continuously changing visual information, such as videos and live streaming.

[0102] This invention functions as a real-time visual information assistance system for content distribution services. Users use this system on a smart device to view streaming content. The terminal continuously collects visual information and transmits it to a central processing unit for analysis. The central processing unit utilizes artificial intelligence models to analyze images and videos. Its main software technologies include video segmentation using OpenCV and analysis using deep learning models with PyTorch.

[0103] The analyzed data is transformed into a textual description using a generative AI model. This textual description is generated in real time in sync with the streaming and sent back to the device. Finally, the device outputs this textual description as speech using speech synthesis technology, such as the Google Text-to-Speech API, conveying the content to the user in real time.

[0104] As a concrete example, the system analyzes in real time a scene in a movie where a bus suddenly stops in front of a character while the user is watching. An example of a prompt for the generative model would be: "Analyze the movie scene and describe the main characters and their actions in detail. Example: 'A woman in a red dress is standing at the bus stop.' Cover all important actions." This allows visually impaired users to perceive the content of the movie as if they were actually seeing the scene.

[0105] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0106] Step 1:

[0107] The device collects visual information from streaming video data by dividing it into frames. Specifically, it uses software such as OpenCV to capture 30 frames per second and temporarily stores this visual information. The input is streaming video data, and the output is image information divided into frames.

[0108] Step 2:

[0109] The terminal sends the collected visual information as a data request to the central processing unit. Here, each frame is processed as an API request and sent to the server. The input is segmented image information, and the output is the data request sent via the API.

[0110] Step 3:

[0111] The server analyzes the received visual information using an artificial intelligence model. Specifically, it uses a PyTorch-based object recognition model to detect objects and people within each frame. The input is the transmitted frame image, and the output is the detected object information.

[0112] Step 4:

[0113] The server uses a generative AI model to generate text descriptions based on the analysis results. Here, models such as GPT-3(registered trademark) 5 are used to create captions in natural language based on the input object information. The input is object information, and the output is the generated text description.

[0114] Step 5:

[0115] The server sends the generated text description to the terminal. Specifically, it returns text data to the terminal as a data response. The input is the generated text description, and the output is the text data sent to the terminal.

[0116] Step 6:

[0117] The device outputs the received text description as audio. Specifically, it uses the Google Text-to-Speech API to play the caption as synthesized speech. The input is the text description received from the server, and the output is audio data that the user can listen to.

[0118] Through the processing steps described above, visually impaired individuals can grasp visual information in real time and obtain details of dynamic content via audio.

[0119] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0120] This invention provides a system for visually impaired individuals to understand image content via the internet and social networking services, and also incorporates an emotion engine for generating captions that take user emotions into account. This system includes a terminal, a server, and a speech-generating function, as well as an emotion engine with emotion recognition capabilities.

[0121] Users use their devices to browse websites and social media, accessing pages that interest them. The devices automatically collect image data from these pages and send it to the server.

[0122] The server performs image analysis on the received image data using an AI model to analyze the image content in detail. This analysis process utilizes functions such as object recognition and scene detection. Based on the analysis results, the server generates a caption describing the image content.

[0123] In addition, the server uses data from the terminal and an emotion engine to recognize the user's emotions based on their voice responses and behavioral data. This recognized emotion information is used to adjust the tone and content of the generated captions. For example, if the image is a landscape photograph and the user is recognized as being in a relaxed state, the caption may be expressed in a more emotional tone.

[0124] The generated captions are sent to the device, converted into speech using speech synthesis technology, and then provided to the user. Users can listen to the captions as audio, and because appropriate content adjustments are made based on emotion, they can enjoy a richer informational experience.

[0125] In this way, systems equipped with emotion recognition capabilities go beyond simply providing information; they enable flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0126] The following describes the processing flow.

[0127] Step 1:

[0128] A user uses their device to access a website or social networking service and opens a page. During this process, the user performs the usual actions.

[0129] Step 2:

[0130] The device detects images displayed on the open page and retrieves the image data. The images are stored in digital format.

[0131] Step 3:

[0132] The device sends the acquired image data to the server. This transmission is usually done via an API request.

[0133] Step 4:

[0134] The server inputs the received image data into an AI model and analyzes the image content. This analysis includes object recognition and scene identification.

[0135] Step 5:

[0136] The server generates image captions based on the analysis results. These captions are written in natural language.

[0137] Step 6:

[0138] The server uses an emotion engine based on audio and behavioral data received from the terminal to recognize the user's emotions. This emotion data is then used to generate captions.

[0139] Step 7:

[0140] The server adjusts the caption's wording, customizing its tone and content based on the recognized emotion.

[0141] Step 8:

[0142] The adjusted caption is sent back to the terminal. The terminal receives the caption and converts it into speech through a speech synthesis engine.

[0143] Step 9:

[0144] Users understand the content of images by listening to audio captions from their devices. Emotion-based tone adjustments provide a more comfortable information experience.

[0145] (Example 2)

[0146] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0147] Technologies for providing easily understandable image content on the internet to visually impaired individuals are limited, and there is a particular need to provide information that considers the user's emotions in addition to understanding the images. However, conventional systems simply convert image content into audio, which is insufficient in providing information that meets the user's emotional needs. As a result, recipients of information do not have a meaningful experience based on their emotions, which is a problem.

[0148] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0149] In this invention, the server includes means for analyzing image data using AI technology to identify objects and detect scenes, means for generating captions based on the analysis results using generative AI technology, and means for adjusting captions considering the user's emotional state. This enables detailed interpretation of images as well as the provision of information tailored to the user's emotions, thereby providing an emotionally valuable information experience for visually impaired individuals.

[0150] A "terminal" is a device used by users to access image content on the internet and social networks and to collect that data.

[0151] A "server" is a central system that receives image data transmitted from terminals, performs analysis, and processes data based on the results.

[0152] "Image data" refers to a collection of digital data that represents visual information existing on the internet or social networks.

[0153] "AI technology" refers to advanced information processing technology used to analyze image data and perform tasks such as object identification and scene detection.

[0154] "Generative AI technology" is a technology for generating natural language captions from analyzed image data.

[0155] A "caption" is a short sentence generated to describe the content of an image in language.

[0156] "User emotional state" refers to information that indicates the psychological reaction and circumstances of a user when they receive an image.

[0157] "Speech synthesis technology" is a technology that converts text data into speech data, and is a means of providing users with auditory information.

[0158] This invention provides a system for visually impaired individuals to understand image content on the internet and social networks. This system enables flexible information delivery, including a caption generation function that takes user emotions into consideration.

[0159] The user first uses their device to browse websites and social networking services through an internet browser or dedicated application. The device automatically detects and collects image data provided on these web pages. The collected image data is sent to the server using a secure and reliable communication protocol (e.g., HTTPS).

[0160] The server is located on hardware designed to analyze received image data and utilizes high-performance deep learning frameworks (e.g., TensorFlow, PyTorch) as its AI technology. The server leverages techniques such as Convolutional Neural Networks (CNNs) to perform object identification and scene detection, enabling a detailed understanding of the image content.

[0161] Based on the analyzed results, the server generates captions using generative AI technology (e.g., natural language processing models). Prompts are used in this process to instruct the model, organizing the information and generating clear sentences. An example of a prompt might be, "Describe in detail what is depicted in this image."

[0162] Furthermore, the server uses the user's voice data and behavioral logs collected from the terminal to estimate the user's emotional state using an emotion engine. Through speech recognition technology (e.g., speech-to-text services), it analyzes the user's dialogue and tone to evaluate their emotions. Based on the recognized emotion data, it adjusts the tone and content of the generated captions to enable expressions that better reflect the user's emotions.

[0163] Finally, the adjusted captions are sent to the device and converted into speech using speech synthesis technology (e.g., a speech conversion service). By listening to this, users can properly understand the visual information and enjoy a deeper informational experience.

[0164] This system goes beyond simply providing information; it enables flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0165] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0166] Step 1:

[0167] Users browse the internet and social networking pages through their devices. The devices automatically detect and collect image data displayed on these pages. Input includes web page URLs and HTML data via browsers and applications, and output is image data. The devices temporarily store the image data to prepare it for subsequent processing.

[0168] Step 2:

[0169] The terminal sends the collected image data to the server. The input is the image data collected in step 1, and the output is the delivery of the image data to the server. The HTTPS protocol is used for transmission to securely transfer the data. This process prepares the image data for analysis.

[0170] Step 3:

[0171] The server receives transmitted image data and performs analysis using AI technology. The input is image data sent from the terminal, and the output is the analysis results from object recognition and scene detection. As for AI technology, a deep learning framework (e.g., TensorFlow or PyTorch) is used, and CNNs are utilized to identify objects and scenes in the image. This process allows for a detailed understanding of the image content.

[0172] Step 4:

[0173] The server uses the analysis results to generate captions based on generative AI technology. The input is the analysis results obtained in step 3, and the output is a caption that describes the object or scene. The generative AI model uses natural language processing technology, and appropriate captions are generated by providing prompts. For example, a prompt such as "Describe what is in this image" will produce a detailed caption of the image content.

[0174] Step 5:

[0175] The server analyzes user voice data and behavior logs provided by the terminal and evaluates the user's emotions using an emotion engine. The input is voice data and operation logs, and the output is the evaluation result of the user's emotional state. Emotions are inferred from voice tone and operation tendencies based on speech recognition technology and behavior pattern analysis.

[0176] Step 6:

[0177] The server adjusts the caption based on the recognized user's emotions. The input is the caption generated in step 4 and the emotion data evaluated in step 5, and the output is the adjusted caption that matches the emotion. The caption may be changed, for example, to a relaxed tone, to provide information that matches the user's emotional state.

[0178] Step 7:

[0179] The server sends the adjusted caption to the terminal. The input is the adjusted caption obtained in step 6, and the output is the transfer of the data containing this caption to the terminal.

[0180] Step 8:

[0181] The device converts received captions into speech using speech synthesis technology. The input is the caption sent from the server, and the output is audio data that the user can listen to. Speech synthesis technology converts text into natural-sounding speech, which is then provided to the user. As a result, users can understand image information on the internet aurally and experience emotionally adapted content.

[0182] (Application Example 2)

[0183] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0184] Visually impaired individuals have difficulty effectively understanding video content on the internet. To address this problem, it is necessary to go beyond simply providing information and adjust the content according to the user's emotional state. While conventional technologies can generate captions that substitute for information the user receives visually, a challenge remains in the lack of a mechanism to dynamically adjust the content in response to the user's emotions.

[0185] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0186] In this invention, the server includes means for receiving and analyzing video data from a terminal, means for analyzing the user's voice response and recognizing emotions, and means for dynamically adjusting the content of caption generation based on the recognized emotions. This allows visually impaired individuals to obtain information appropriate to their emotional state, enabling a more empathetic understanding of the content.

[0187] A "terminal" is an information processing device used by a user, and is a device that has the function of collecting image and audio data and running applications.

[0188] "Motion image data" refers to visual data composed of a series of frames, and includes information in the form of videos and animations.

[0189] A "server" is a centralized computer that receives data from terminals, processes and analyzes it, and provides services to other devices on a communication network.

[0190] "Analysis" is the process of identifying specific patterns or features in received data and deciphering its content.

[0191] A "caption" is a descriptive text generated based on video and image data, representing visual information as textual information.

[0192] "Speech synthesis" is a technology that converts text data into speech, and is a process in which a computer outputs it as spoken language.

[0193] "Voice response" refers to the tone of voice and words spoken by the user through the device, and is information that indicates the user's emotions and state of mind.

[0194] "Emotion recognition" is the process of inferring and identifying a user's emotional state from their voice responses and other data.

[0195] "Adjustment" is the process of modifying the generated content or its presentation according to specific conditions or circumstances.

[0196] This invention provides a system for visually impaired individuals to effectively understand video content and enables flexible caption generation that responds to the user's emotional state. Specific embodiments are described below.

[0197] Suitable devices for users include smartphones and head-mounted displays, which have built-in cameras and microphones. These devices access content via the internet and send video and image data of interest to a server. This transmission is done through an application programming interface.

[0198] The server uses generative AI models for image analysis (e.g., YOLO or ResNet) to analyze the received video data. During this process, object recognition and scene detection are performed. Based on the analysis results, captions describing the content of the video are generated. Furthermore, the server analyzes the user's voice responses from the terminal and uses an emotion recognition engine (e.g., OpenAI®'s GPT-3) to identify the user's emotional state. This emotion information is used to adjust the content and tone of the generated captions.

[0199] The generated captions are converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and provided to the user. This allows users to complement visual information through emotionally appropriate audio content.

[0200] For example, if a user is watching an action scene in a movie, the emotion recognition results will generate a caption such as "An exciting scene is unfolding." On the other hand, for a quiet, moving scene, the caption will be adjusted to "Heartwarming music is playing." An example of this prompt would be, "Please describe the content of this video scene as an audio guide. Generate a caption while describing the content, using a calm tone if the user is calm, and emphasizing if they are excited."

[0201] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0202] Step 1:

[0203] The terminal retrieves video and image data of interest to the user via the internet and sends it to the server via an application programming interface. The input is video and image data, and the output is data transmission to the server. In this process, the terminal appropriately formats the video and image data and transmits it as packets over the network.

[0204] Step 2:

[0205] The server acquires video data received from the terminal and begins analyzing the data using a generating AI model. The input is video data received from the terminal, and the output is information on object recognition and scene detection as a result of the analysis. The server uses an AI model (e.g., YOLO or ResNet) to identify features of objects and scenes in the video and records this information in a database.

[0206] Step 3:

[0207] The server generates a caption describing the content of the video based on the analysis results. The input is the analysis results from step 2, and the output is the caption text. Natural language generation technology using prompt sentences is applied to generate the caption, expressing the analyzed information in language that is easy for the user to understand.

[0208] Step 4:

[0209] The device collects the user's voice responses and sends them to the server. The input is the user's voice, and the output is the transmission of voice data to the server. The device uses a microphone to capture the voice in real time, compresses the data as needed, and sends it to the server.

[0210] Step 5:

[0211] The server uses a speech recognition engine to analyze the user's emotions from their voice data. The input is the voice data from step 4, and the output is the analyzed user emotion information. The server analyzes the voice waveform and estimates the user's emotional state using template matching and machine learning.

[0212] Step 6:

[0213] The server uses the analyzed sentiment information to adjust the tone and content of the caption. The input is the caption text from step 3 and the sentiment information from step 5, and the output is the adjusted caption. Depending on the sentiment recognition results, the intonation and expression of the caption change, and the final audio output is generated.

[0214] Step 7:

[0215] Finally, the captions generated by the server are converted into speech using speech synthesis technology and sent to the terminal. The input is the adjusted caption text, and the output is audio data. The terminal plays this audio and provides it to the user. Existing technologies such as Google Cloud Text-to-Speech are used for speech synthesis to reproduce natural-sounding speech.

[0216] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0217] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0218] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0219] [Second Embodiment]

[0220] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0221] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0222] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0223] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0224] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0225] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0226] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0227] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0228] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0229] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0230] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0231] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. 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".

[0232] This invention provides a system that enables visually impaired individuals to understand image content on the internet and social networking services. The system includes a terminal, a server, and software with caption generation and speech conversion capabilities.

[0233] Users typically use their devices to navigate websites and social media, opening pages of interest. During this process, the device has a function to automatically detect images on the page. The detected image data is then transmitted to a server via the internet.

[0234] The server uses an AI model to analyze the received image data. This AI model has the ability to analyze the image content in detail using technologies such as object recognition and scene detection. Once the analysis is complete, the server generates text captions corresponding to the images.

[0235] The generated captions are sent back to the device and converted into audio. This conversion uses natural-sounding speech synthesis technology, allowing users to hear the content of the image in audio form.

[0236] For example, if a user opens a travel blog page and tries to view a photograph of a mountain landscape, the image is sent to the server, where a caption such as "The Alps and lush green valleys stretching out under a blue sky" is generated. The device then provides this caption to the user via audio, allowing them to understand the image's content without relying on visual information.

[0237] This system provides a barrier-free environment for visually impaired individuals, enabling them to access images without alternative text, and significantly improving their online information access.

[0238] The following describes the processing flow.

[0239] Step 1:

[0240] A user uses their device to access websites and social networking services and opens pages of interest. During this process, the user performs typical browsing actions.

[0241] Step 2:

[0242] The device automatically detects images on the opened page and temporarily retrieves the image data. The images are stored in URL or base64 format.

[0243] Step 3:

[0244] The terminal prepares to send the acquired image data to the server and sends the image data to the server via an internet connection in the form of an API request.

[0245] Step 4:

[0246] The server receives image data sent from the terminal and begins analyzing the image using an AI model. This analysis includes object recognition and scene detection.

[0247] Step 5:

[0248] The server generates image captions based on the analysis results. The generated captions describe the details of the objects and scenes in the image using natural language.

[0249] Step 6:

[0250] The server sends the generated caption back to the terminal. At this time, the caption is sent in a standard data format.

[0251] Step 7:

[0252] The terminal receives captions from the server and uses a text-to-speech engine to convert the text captions into speech.

[0253] Step 8:

[0254] The device provides the user with an audio caption via a speaker or headphones, allowing the user to understand the image content audibly.

[0255] (Example 1)

[0256] Next, we will describe Example 1. 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."

[0257] Individuals with visual impairments face difficulties in effectively understanding visual content on the internet and digital media. Furthermore, systems providing audio information as an alternative to visual information are not sufficiently developed, resulting in a lack of means to accurately convey the meaning of visual content. This situation leads to the challenge of limited information access for individuals with visual impairments.

[0258] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0259] In this invention, the server includes means for analyzing visual information and generating natural language descriptions, means for a mechanism for analyzing visual signals, and means for using sound generation technology to generate naturally sounding synthesized speech. This enables individuals with visual impairments to accurately understand visual information as sound and to smoothly access information on the internet and digital media.

[0260] A "terminal" is a computer device used to detect visual information and transmit it to an information processing device.

[0261] "Visual information" refers to visual data detected from a device as digital images or videos.

[0262] An "information processing device" is a computer device that analyzes visual information received from a terminal and generates natural language descriptions.

[0263] "Natural language description" refers to a form of text that is understandable to humans and is generated based on visual information.

[0264] "Visual signals" refer to image and video data that a device detects and analyzes.

[0265] "Sound generation technology" is a technology that converts text data into speech that sounds like human speech.

[0266] A "learning model" is a computational model that analyzes visual information based on AI technology to perform feature recognition and scene detection.

[0267] This invention provides a system for individuals with visual impairments to understand visual information on the internet and digital media as audio. The system includes a terminal, an information processing device, and software employing sound generation technology to convert visual information into audio.

[0268] Users access web pages and social media content containing visual information through their devices. These devices have the capability to detect visual information on the page and transmit this information to an information processing device. The devices used are commonly available personal computers and smart devices.

[0269] The server, as an information processing unit, is equipped with a learning model for analyzing received visual information. This model is implemented using widely used AI frameworks such as TensorFlow and PyTorch, and recognizes specific features from visual information to generate natural language descriptions. For example, it performs object recognition and scene detection, and then expresses the visual information as text based on the results.

[0270] The generated natural language description is then sent back to the device. The device uses audio generation technologies, such as the Google Text-to-Speech API, to convert the natural language description into synthesized speech. This allows the user to understand visual information through speech.

[0271] As a concrete example, consider a case where a user views a blog post containing a photograph of a tourist destination. The device sends this photograph to the server, which generates a natural language description from the visual information, such as "a historical castle and its surrounding landscape against a blue sky." The device then provides this description to the user as synthesized speech. An example of a prompt could be the instruction, "Please create a caption that describes the content of this image in detail."

[0272] By implementing this invention, individuals with visual impairments can receive visual information as audio, significantly improving their access to information provided online.

[0273] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0274] Step 1:

[0275] The user opens internet or social media pages on their device. The device automatically detects images contained in the page. Specifically, it analyzes the HTML and CSS, This tool extracts data set as background images using tags and CSS. The input is the HTML data of a webpage, and the output is the image URL or binary data.

[0276] Step 2:

[0277] The device sends the detected image data to the server. Specifically, it sends the image data using an HTTP POST request. In this process, the image metadata may also be sent. The input is image data (e.g., JPEG or PNG format), and the output is an HTTP request that is sent to the server.

[0278] Step 3:

[0279] The server inputs the received image data into an AI model. The AI ​​model is, for example, a deep learning model using TensorFlow, and performs object recognition and scene detection. The input is image data, and the output is intermediate data describing the image's features. Specifically, a convolutional neural network (CNN) extracts the main features of the image.

[0280] Step 4:

[0281] The server generates a natural language description based on the output of the AI model. The generative AI model analyzes the feature data and generates the corresponding text. For example, it creates a caption such as "A historical castle spreading under the blue sky". The input is the feature data of the AI model, and the output is the natural language description.

[0282] Step 5:

[0283] The server sends the generated natural language description to the terminal. It returns an HTTP response to transmit the natural language description to the terminal. The input is the generated natural language description, and the output is the text data sent to the terminal.

[0284] Step 6:

[0285] The terminal inputs the received natural language description into the speech synthesis engine to generate speech. For example, it uses the Google Text-to-Speech API to convert text into speech. The input is the natural language text, and the output is the synthesized speech data. As a specific operation, it converts the text into phonemes and generates a sound waveform.

[0286] Step 7:

[0287] The user listens to the speech provided by the terminal. Thereby, the user can grasp the visual information in the form of speech. The terminal plays the speech via a speaker or earphone. This process includes operations such as playback control and volume adjustment. The input is the synthesized speech data, and the output is the speech played to the user.

[0288] (Application Example 1)

[0289] Next, Application Example 1 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart glasses 214 are referred to as the "terminal".

[0290] Visually impaired individuals face the challenge of understanding online video and image content in real time. Furthermore, general visual caption generation systems are specialized for static images only and are insufficient to assist in understanding dynamic content. Specifically, there is a need for methods to grasp the actions of characters and details of scenes in video works such as movies and television programs in real time, without relying on sight.

[0291] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0292] In this invention, the server includes means for collecting visual information from a terminal and transmitting it to a central processing unit, means for the central processing unit to analyze the visual information and generate a text description, and means for transmitting the generated text description to the terminal and outputting it as audio. This makes it possible for visually impaired people to understand visual information in real time and to enrich their experience of dynamic content.

[0293] A "terminal" is a device that collects visual information and transmits it to a central processing unit.

[0294] "Visual information" refers to visual content such as images and videos, which are analyzed by a central processing unit.

[0295] A "central processing unit" is a system that analyzes visual information transmitted from terminals and generates textual descriptions.

[0296] A "text description" is a written representation of the content of the analyzed visual information, generated to help visually impaired individuals understand its content.

[0297] A "generative model" refers to artificial intelligence technology used to generate appropriate text descriptions from visual information.

[0298] "Audio output" refers to the process of providing the generated text explanation to the user as audio.

[0299] "Real-time" refers to a state where events are processed simultaneously, meaning that information is provided without delay.

[0300] "Dynamic content" refers to content that includes continuously changing visual information, such as videos and live streaming.

[0301] This invention functions as a real-time visual information assistance system for content distribution services. Users use this system on a smart device to view streaming content. The terminal continuously collects visual information and transmits it to a central processing unit for analysis. The central processing unit utilizes artificial intelligence models to analyze images and videos. Its main software technologies include video segmentation using OpenCV and analysis using deep learning models with PyTorch.

[0302] The analyzed data is transformed into a textual description using a generative AI model. This textual description is generated in real time in sync with the streaming and sent back to the device. Finally, the device outputs this textual description as speech using speech synthesis technology, such as the Google Text-to-Speech API, conveying the content to the user in real time.

[0303] As a concrete example, the system analyzes in real time a scene in a movie where a bus suddenly stops in front of a character while the user is watching. An example of a prompt for the generative model would be: "Analyze the movie scene and describe the main characters and their actions in detail. Example: 'A woman in a red dress is standing at the bus stop.' Cover all important actions." This allows visually impaired users to perceive the content of the movie as if they were actually seeing the scene.

[0304] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0305] Step 1:

[0306] The terminal divides and collects visual information from the video data during streaming. As a specific operation, software such as OpenCV is used to capture at 30 frames per second, and those visual information are temporarily saved. The input is the streaming video data, and the output is the image information divided frame by frame.

[0307] Step 2:

[0308] The terminal sends the collected visual information to the central processing unit as a data request. Here, each frame is processed as an API request and sent to the server. The input is the divided image information, and the output is the data request sent via the API.

[0309] Step 3:

[0310] The server analyzes the received visual information using an artificial intelligence model. As a specific operation, a PyTorch-based object recognition model is used to detect objects and people within each frame. The input is the transmitted frame image, and the output is the detected object information.

[0311] Step 4:

[0312] The server generates a text description based on the analysis results using a generative AI model. Here, a model such as GPT-3.5 is used to create a caption in natural language based on the input object information. The input is the object information, and the output is the generated text description.

[0313] Step 5:

[0314] The server sends the generated text description to the terminal. As a specific operation, the text data is returned to the terminal as a data response. The input is the generated text description, and the output is the text data sent to the terminal.

[0315] Step 6:

[0316] The device outputs the received text description as audio. Specifically, it uses the Google Text-to-Speech API to play the caption as synthesized speech. The input is the text description received from the server, and the output is audio data that the user can listen to.

[0317] Through the processing steps described above, visually impaired individuals can grasp visual information in real time and obtain details of dynamic content via audio.

[0318] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0319] This invention provides a system for visually impaired individuals to understand image content via the internet and social networking services, and also incorporates an emotion engine for generating captions that take user emotions into account. This system includes a terminal, a server, and a speech-generating function, as well as an emotion engine with emotion recognition capabilities.

[0320] Users use their devices to browse websites and social media, accessing pages that interest them. The devices automatically collect image data from these pages and send it to the server.

[0321] The server performs image analysis on the received image data using an AI model to analyze the image content in detail. This analysis process utilizes functions such as object recognition and scene detection. Based on the analysis results, the server generates a caption describing the image content.

[0322] In addition, the server uses data from the terminal and an emotion engine to recognize the user's emotions based on their voice responses and behavioral data. This recognized emotion information is used to adjust the tone and content of the generated captions. For example, if the image is a landscape photograph and the user is recognized as being in a relaxed state, the caption may be expressed in a more emotional tone.

[0323] The generated captions are sent to the device, converted into speech using speech synthesis technology, and then provided to the user. Users can listen to the captions as audio, and because appropriate content adjustments are made based on emotion, they can enjoy a richer informational experience.

[0324] In this way, systems equipped with emotion recognition capabilities go beyond simply providing information; they enable flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0325] The following describes the processing flow.

[0326] Step 1:

[0327] A user uses their device to access a website or social networking service and opens a page. During this process, the user performs the usual actions.

[0328] Step 2:

[0329] The device detects images displayed on the open page and retrieves the image data. The images are stored in digital format.

[0330] Step 3:

[0331] The device sends the acquired image data to the server. This transmission is usually done via an API request.

[0332] Step 4:

[0333] The server inputs the received image data into an AI model and analyzes the image content. This analysis includes object recognition and scene identification.

[0334] Step 5:

[0335] The server generates image captions based on the analysis results. These captions are written in natural language.

[0336] Step 6:

[0337] The server uses an emotion engine based on audio and behavioral data received from the terminal to recognize the user's emotions. This emotion data is then used to generate captions.

[0338] Step 7:

[0339] The server adjusts the caption's wording, customizing its tone and content based on the recognized emotion.

[0340] Step 8:

[0341] The adjusted caption is sent back to the terminal. The terminal receives the caption and converts it into speech through a speech synthesis engine.

[0342] Step 9:

[0343] Users understand the content of images by listening to audio captions from their devices. Emotion-based tone adjustments provide a more comfortable information experience.

[0344] (Example 2)

[0345] Next, we will describe Example 2. 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".

[0346] Technologies for providing easily understandable image content on the internet to visually impaired individuals are limited, and there is a particular need to provide information that considers the user's emotions in addition to understanding the images. However, conventional systems simply convert image content into audio, which is insufficient in providing information that meets the user's emotional needs. As a result, recipients of information do not have a meaningful experience based on their emotions, which is a problem.

[0347] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0348] In this invention, the server includes means for analyzing image data using AI technology to identify objects and detect scenes, means for generating captions based on the analysis results using generative AI technology, and means for adjusting captions considering the user's emotional state. This enables detailed interpretation of images as well as the provision of information tailored to the user's emotions, thereby providing an emotionally valuable information experience for visually impaired individuals.

[0349] A "terminal" is a device used by users to access image content on the internet and social networks and to collect that data.

[0350] A "server" is a central system that receives image data transmitted from terminals, performs analysis, and processes data based on the results.

[0351] "Image data" refers to a collection of digital data that represents visual information existing on the internet or social networks.

[0352] "AI technology" refers to advanced information processing technology used to analyze image data and perform tasks such as object identification and scene detection.

[0353] "Generative AI technology" is a technology for generating natural language captions from analyzed image data.

[0354] A "caption" is a short sentence generated to describe the content of an image in language.

[0355] "User emotional state" refers to information that indicates the psychological reaction and circumstances of a user when they receive an image.

[0356] "Speech synthesis technology" is a technology that converts text data into speech data, and is a means of providing users with auditory information.

[0357] This invention provides a system for visually impaired individuals to understand image content on the internet and social networks. This system enables flexible information delivery, including a caption generation function that takes user emotions into consideration.

[0358] The user first uses their device to browse websites and social networking services through an internet browser or dedicated application. The device automatically detects and collects image data provided on these web pages. The collected image data is sent to the server using a secure and reliable communication protocol (e.g., HTTPS).

[0359] The server is located on hardware designed to analyze received image data and utilizes high-performance deep learning frameworks (e.g., TensorFlow, PyTorch) as its AI technology. The server leverages techniques such as Convolutional Neural Networks (CNNs) to perform object identification and scene detection, enabling a detailed understanding of the image content.

[0360] Based on the analyzed results, the server generates captions using generative AI technology (e.g., natural language processing models). Prompts are used in this process to instruct the model, organizing the information and generating clear sentences. An example of a prompt might be, "Describe in detail what is depicted in this image."

[0361] Furthermore, the server uses the user's voice data and behavioral logs collected from the terminal to estimate the user's emotional state using an emotion engine. Through speech recognition technology (e.g., speech-to-text services), it analyzes the user's dialogue and tone to evaluate their emotions. Based on the recognized emotion data, it adjusts the tone and content of the generated captions to enable expressions that better reflect the user's emotions.

[0362] Finally, the adjusted captions are sent to the device and converted into speech using speech synthesis technology (e.g., a speech conversion service). By listening to this, users can properly understand the visual information and enjoy a deeper informational experience.

[0363] This system goes beyond simply providing information; it enables flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0364] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0365] Step 1:

[0366] Users browse the internet and social networking pages through their devices. The devices automatically detect and collect image data displayed on these pages. Input includes web page URLs and HTML data via browsers and applications, and output is image data. The devices temporarily store the image data to prepare it for subsequent processing.

[0367] Step 2:

[0368] The terminal sends the collected image data to the server. The input is the image data collected in step 1, and the output is the delivery of the image data to the server. The HTTPS protocol is used for transmission to securely transfer the data. This process prepares the image data for analysis.

[0369] Step 3:

[0370] The server receives transmitted image data and performs analysis using AI technology. The input is image data sent from the terminal, and the output is the analysis results from object recognition and scene detection. As for AI technology, a deep learning framework (e.g., TensorFlow or PyTorch) is used, and CNNs are utilized to identify objects and scenes in the image. This process allows for a detailed understanding of the image content.

[0371] Step 4:

[0372] The server uses the analysis results to generate captions based on generative AI technology. The input is the analysis results obtained in step 3, and the output is a caption that describes the object or scene. The generative AI model uses natural language processing technology, and appropriate captions are generated by providing prompts. For example, a prompt such as "Describe what is in this image" will produce a detailed caption of the image content.

[0373] Step 5:

[0374] The server analyzes user voice data and behavior logs provided by the terminal and evaluates the user's emotions using an emotion engine. The input is voice data and operation logs, and the output is the evaluation result of the user's emotional state. Emotions are inferred from voice tone and operation tendencies based on speech recognition technology and behavior pattern analysis.

[0375] Step 6:

[0376] The server adjusts the caption based on the recognized user's emotions. The input is the caption generated in step 4 and the emotion data evaluated in step 5, and the output is the adjusted caption that matches the emotion. The caption may be changed, for example, to a relaxed tone, to provide information that matches the user's emotional state.

[0377] Step 7:

[0378] The server sends the adjusted caption to the terminal. The input is the adjusted caption obtained in step 6, and the output is the transfer of the data containing this caption to the terminal.

[0379] Step 8:

[0380] The device converts received captions into speech using speech synthesis technology. The input is the caption sent from the server, and the output is audio data that the user can listen to. Speech synthesis technology converts text into natural-sounding speech, which is then provided to the user. As a result, users can understand image information on the internet aurally and experience emotionally adapted content.

[0381] (Application Example 2)

[0382] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0383] Visually impaired individuals have difficulty effectively understanding video content on the internet. To address this problem, it is necessary to go beyond simply providing information and adjust the content according to the user's emotional state. While conventional technologies can generate captions that substitute for information the user receives visually, a challenge remains in the lack of a mechanism to dynamically adjust the content in response to the user's emotions.

[0384] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0385] In this invention, the server includes means for receiving and analyzing video data from a terminal, means for analyzing the user's voice response and recognizing emotions, and means for dynamically adjusting the content of caption generation based on the recognized emotions. This allows visually impaired individuals to obtain information appropriate to their emotional state, enabling a more empathetic understanding of the content.

[0386] A "terminal" is an information processing device used by a user, and is a device that has the function of collecting image and audio data and running applications.

[0387] "Motion image data" refers to visual data composed of a series of frames, and includes information in the form of videos and animations.

[0388] A "server" is a centralized computer that receives data from terminals, processes and analyzes it, and provides services to other devices on a communication network.

[0389] "Analysis" is the process of identifying specific patterns or features in received data and deciphering its content.

[0390] A "caption" is a descriptive text generated based on video and image data, representing visual information as textual information.

[0391] "Speech synthesis" is a technology that converts text data into speech, and is a process in which a computer outputs it as spoken language.

[0392] "Voice response" refers to the tone of voice and words spoken by the user through the device, and is information that indicates the user's emotions and state of mind.

[0393] "Emotion recognition" is the process of inferring and identifying a user's emotional state from their voice responses and other data.

[0394] "Adjustment" is the process of modifying the generated content or its presentation according to specific conditions or circumstances.

[0395] This invention provides a system for visually impaired individuals to effectively understand video content and enables flexible caption generation that responds to the user's emotional state. Specific embodiments are described below.

[0396] Suitable devices for users include smartphones and head-mounted displays, which have built-in cameras and microphones. These devices access content via the internet and send video and image data of interest to a server. This transmission is done through an application programming interface.

[0397] The server uses generative AI models for image analysis (e.g., YOLO or ResNet) to analyze the received video data. During this process, object recognition and scene detection are performed. Based on the analysis results, captions describing the content of the video are generated. Furthermore, the server analyzes the user's voice responses from the terminal and uses an emotion recognition engine (e.g., OpenAI's GPT-3) to identify the user's emotional state. This emotion information is used to adjust the content and tone of the generated captions.

[0398] The generated captions are converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and provided to the user. This allows users to complement visual information through emotionally appropriate audio content.

[0399] For example, if a user is watching an action scene in a movie, the emotion recognition results will generate a caption such as "An exciting scene is unfolding." On the other hand, for a quiet, moving scene, the caption will be adjusted to "Heartwarming music is playing." An example of this prompt would be, "Please describe the content of this video scene as an audio guide. Generate a caption while describing the content, using a calm tone if the user is calm, and emphasizing if they are excited."

[0400] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0401] Step 1:

[0402] The terminal retrieves video and image data of interest to the user via the internet and sends it to the server via an application programming interface. The input is video and image data, and the output is data transmission to the server. In this process, the terminal appropriately formats the video and image data and transmits it as packets over the network.

[0403] Step 2:

[0404] The server acquires video data received from the terminal and begins analyzing the data using a generating AI model. The input is video data received from the terminal, and the output is information on object recognition and scene detection as a result of the analysis. The server uses an AI model (e.g., YOLO or ResNet) to identify features of objects and scenes in the video and records this information in a database.

[0405] Step 3:

[0406] The server generates a caption describing the content of the video based on the analysis results. The input is the analysis results from step 2, and the output is the caption text. Natural language generation technology using prompt sentences is applied to generate the caption, expressing the analyzed information in language that is easy for the user to understand.

[0407] Step 4:

[0408] The device collects the user's voice responses and sends them to the server. The input is the user's voice, and the output is the transmission of voice data to the server. The device uses a microphone to capture the voice in real time, compresses the data as needed, and sends it to the server.

[0409] Step 5:

[0410] The server uses a speech recognition engine to analyze the user's emotions from their voice data. The input is the voice data from step 4, and the output is the analyzed user emotion information. The server analyzes the voice waveform and estimates the user's emotional state using template matching and machine learning.

[0411] Step 6:

[0412] The server uses the analyzed sentiment information to adjust the tone and content of the caption. The input is the caption text from step 3 and the sentiment information from step 5, and the output is the adjusted caption. Depending on the sentiment recognition results, the intonation and expression of the caption change, and the final audio output is generated.

[0413] Step 7:

[0414] Finally, the captions generated by the server are converted into speech using speech synthesis technology and sent to the terminal. The input is the adjusted caption text, and the output is audio data. The terminal plays this audio and provides it to the user. Existing technologies such as Google Cloud Text-to-Speech are used for speech synthesis to reproduce natural-sounding speech.

[0415] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0416] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0417] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0418] [Third Embodiment]

[0419] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0420] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0421] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0422] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0423] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0424] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0425] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0426] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0427] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0428] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0429] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0430] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0431] This invention provides a system that enables visually impaired individuals to understand image content on the internet and social networking services. The system includes a terminal, a server, and software with caption generation and speech conversion capabilities.

[0432] Users typically use their devices to navigate websites and social media, opening pages of interest. During this process, the device has a function to automatically detect images on the page. The detected image data is then transmitted to a server via the internet.

[0433] The server uses an AI model to analyze the received image data. This AI model has the ability to analyze the image content in detail using technologies such as object recognition and scene detection. Once the analysis is complete, the server generates text captions corresponding to the images.

[0434] The generated captions are sent back to the device and converted into audio. This conversion uses natural-sounding speech synthesis technology, allowing users to hear the content of the image in audio form.

[0435] For example, if a user opens a travel blog page and tries to view a photograph of a mountain landscape, the image is sent to the server, where a caption such as "The Alps and lush green valleys stretching out under a blue sky" is generated. The device then provides this caption to the user via audio, allowing them to understand the image's content without relying on visual information.

[0436] This system provides a barrier-free environment for visually impaired individuals, enabling them to access images without alternative text, and significantly improving their online information access.

[0437] The following describes the processing flow.

[0438] Step 1:

[0439] A user uses their device to access websites and social networking services and opens pages of interest. During this process, the user performs typical browsing actions.

[0440] Step 2:

[0441] The device automatically detects images on the opened page and temporarily retrieves the image data. The images are stored in URL or base64 format.

[0442] Step 3:

[0443] The terminal prepares to send the acquired image data to the server and sends the image data to the server via an internet connection in the form of an API request.

[0444] Step 4:

[0445] The server receives image data sent from the terminal and begins analyzing the image using an AI model. This analysis includes object recognition and scene detection.

[0446] Step 5:

[0447] The server generates image captions based on the analysis results. The generated captions describe the details of the objects and scenes in the image using natural language.

[0448] Step 6:

[0449] The server sends the generated caption back to the terminal. At this time, the caption is sent in a standard data format.

[0450] Step 7:

[0451] The terminal receives captions from the server and uses a text-to-speech engine to convert the text captions into speech.

[0452] Step 8:

[0453] The device provides the user with an audio caption via a speaker or headphones, allowing the user to understand the image content audibly.

[0454] (Example 1)

[0455] Next, we will describe Example 1. 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."

[0456] Individuals with visual impairments face difficulties in effectively understanding visual content on the internet and digital media. Furthermore, systems providing audio information as an alternative to visual information are not sufficiently developed, resulting in a lack of means to accurately convey the meaning of visual content. This situation leads to the challenge of limited information access for individuals with visual impairments.

[0457] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0458] In this invention, the server includes means for analyzing visual information and generating natural language descriptions, means for a mechanism for analyzing visual signals, and means for using sound generation technology to generate naturally sounding synthesized speech. This enables individuals with visual impairments to accurately understand visual information as sound and to smoothly access information on the internet and digital media.

[0459] A "terminal" is a computer device used to detect visual information and transmit it to an information processing device.

[0460] "Visual information" refers to visual data detected from a device as digital images or videos.

[0461] An "information processing device" is a computer device that analyzes visual information received from a terminal and generates natural language descriptions.

[0462] "Natural language description" refers to a form of text that is understandable to humans and is generated based on visual information.

[0463] "Visual signals" refer to image and video data that a device detects and analyzes.

[0464] "Sound generation technology" is a technology that converts text data into speech that sounds like human speech.

[0465] A "learning model" is a computational model that analyzes visual information based on AI technology to perform feature recognition and scene detection.

[0466] This invention provides a system for individuals with visual impairments to understand visual information on the internet and digital media as audio. The system includes a terminal, an information processing device, and software employing sound generation technology to convert visual information into audio.

[0467] Users access web pages and social media content containing visual information through their devices. These devices have the capability to detect visual information on the page and transmit this information to an information processing device. The devices used are commonly available personal computers and smart devices.

[0468] The server, as an information processing unit, is equipped with a learning model for analyzing received visual information. This model is implemented using widely used AI frameworks such as TensorFlow and PyTorch, and recognizes specific features from visual information to generate natural language descriptions. For example, it performs object recognition and scene detection, and then expresses the visual information as text based on the results.

[0469] The generated natural language description is then sent back to the device. The device uses audio generation technologies, such as the Google Text-to-Speech API, to convert the natural language description into synthesized speech. This allows the user to understand visual information through speech.

[0470] As a concrete example, consider a case where a user views a blog post containing a photograph of a tourist destination. The device sends this photograph to the server, which generates a natural language description from the visual information, such as "a historical castle and its surrounding landscape against a blue sky." The device then provides this description to the user as synthesized speech. An example of a prompt could be the instruction, "Please create a caption that describes the content of this image in detail."

[0471] By implementing this invention, individuals with visual impairments can receive visual information as audio, significantly improving their access to information provided online.

[0472] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0473] Step 1:

[0474] The user opens internet or social media pages on their device. The device automatically detects images contained in the page. Specifically, it analyzes the HTML and CSS, This tool extracts data set as background images using tags and CSS. The input is the HTML data of a webpage, and the output is the image URL or binary data.

[0475] Step 2:

[0476] The device sends the detected image data to the server. Specifically, it sends the image data using an HTTP POST request. In this process, the image metadata may also be sent. The input is image data (e.g., JPEG or PNG format), and the output is an HTTP request that is sent to the server.

[0477] Step 3:

[0478] The server inputs the received image data into an AI model. The AI ​​model is, for example, a deep learning model using TensorFlow, and performs object recognition and scene detection. The input is image data, and the output is intermediate data describing the image's features. Specifically, a convolutional neural network (CNN) extracts the main features of the image.

[0479] Step 4:

[0480] The server generates natural language descriptions based on the output of the AI ​​model. The generative AI model analyzes feature data and generates corresponding text. For example, it might create a caption such as "A historical castle spread out under a blue sky." The input is the feature data of the AI ​​model, and the output is a natural language description.

[0481] Step 5:

[0482] The server sends the generated natural language description to the terminal. It returns an HTTP response, transmitting the natural language description to the terminal. The input is the generated natural language description, and the output is the text data sent to the terminal.

[0483] Step 6:

[0484] The device inputs the received natural language description into a speech synthesis engine and generates speech. For example, it uses the Google Text-to-Speech API to convert text into speech. The input is natural language text, and the output is synthesized speech data. Specifically, it converts the text into phonemes and generates a sound waveform.

[0485] Step 7:

[0486] The user listens to audio provided by the device. This allows the user to perceive visual information audibly. The device plays the audio through speakers or headphones. This process includes playback controls and volume adjustments. The input is synthesized audio data, and the output is the audio played back to the user.

[0487] (Application Example 1)

[0488] Next, we will explain Application Example 1. In the following explanation, 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."

[0489] Visually impaired individuals face the challenge of understanding online video and image content in real time. Furthermore, general visual caption generation systems are specialized for static images only and are insufficient to assist in understanding dynamic content. Specifically, there is a need for methods to grasp the actions of characters and details of scenes in video works such as movies and television programs in real time, without relying on sight.

[0490] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0491] In this invention, the server includes means for collecting visual information from a terminal and transmitting it to a central processing unit, means for the central processing unit to analyze the visual information and generate a text description, and means for transmitting the generated text description to the terminal and outputting it as audio. This makes it possible for visually impaired people to understand visual information in real time and to enrich their experience of dynamic content.

[0492] A "terminal" is a device that collects visual information and transmits it to a central processing unit.

[0493] "Visual information" refers to visual content such as images and videos, which are analyzed by a central processing unit.

[0494] A "central processing unit" is a system that analyzes visual information transmitted from terminals and generates textual descriptions.

[0495] A "text description" is a written representation of the content of the analyzed visual information, generated to help visually impaired individuals understand its content.

[0496] A "generative model" refers to artificial intelligence technology used to generate appropriate text descriptions from visual information.

[0497] "Audio output" refers to the process of providing the generated text explanation to the user as audio.

[0498] "Real-time" refers to a state where events are processed simultaneously, meaning that information is provided without delay.

[0499] "Dynamic content" refers to content that includes continuously changing visual information, such as videos and live streaming.

[0500] This invention functions as a real-time visual information assistance system for content distribution services. Users use this system on a smart device to view streaming content. The terminal continuously collects visual information and transmits it to a central processing unit for analysis. The central processing unit utilizes artificial intelligence models to analyze images and videos. Its main software technologies include video segmentation using OpenCV and analysis using deep learning models with PyTorch.

[0501] The analyzed data is transformed into a textual description using a generative AI model. This textual description is generated in real time in sync with the streaming and sent back to the device. Finally, the device outputs this textual description as speech using speech synthesis technology, such as the Google Text-to-Speech API, conveying the content to the user in real time.

[0502] As a concrete example, the system analyzes in real time a scene in a movie where a bus suddenly stops in front of a character while the user is watching. An example of a prompt for the generative model would be: "Analyze the movie scene and describe the main characters and their actions in detail. Example: 'A woman in a red dress is standing at the bus stop.' Cover all important actions." This allows visually impaired users to perceive the content of the movie as if they were actually seeing the scene.

[0503] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0504] Step 1:

[0505] The device collects visual information from streaming video data by dividing it into frames. Specifically, it uses software such as OpenCV to capture 30 frames per second and temporarily stores this visual information. The input is streaming video data, and the output is image information divided into frames.

[0506] Step 2:

[0507] The terminal sends the collected visual information as a data request to the central processing unit. Here, each frame is processed as an API request and sent to the server. The input is segmented image information, and the output is the data request sent via the API.

[0508] Step 3:

[0509] The server analyzes the received visual information using an artificial intelligence model. Specifically, it uses a PyTorch-based object recognition model to detect objects and people within each frame. The input is the transmitted frame image, and the output is the detected object information.

[0510] Step 4:

[0511] The server uses a generative AI model to generate text descriptions based on the analysis results. Here, a model such as GPT-3.5 is used to create captions in natural language based on the input object information. The input is object information, and the output is the generated text description.

[0512] Step 5:

[0513] The server sends the generated text description to the terminal. Specifically, it returns text data to the terminal as a data response. The input is the generated text description, and the output is the text data sent to the terminal.

[0514] Step 6:

[0515] The device outputs the received text description as audio. Specifically, it uses the Google Text-to-Speech API to play the caption as synthesized speech. The input is the text description received from the server, and the output is audio data that the user can listen to.

[0516] Through the processing steps described above, visually impaired individuals can grasp visual information in real time and obtain details of dynamic content via audio.

[0517] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0518] This invention provides a system for visually impaired individuals to understand image content via the internet and social networking services, and also incorporates an emotion engine for generating captions that take user emotions into account. This system includes a terminal, a server, and a speech-generating function, as well as an emotion engine with emotion recognition capabilities.

[0519] Users use their devices to browse websites and social media, accessing pages that interest them. The devices automatically collect image data from these pages and send it to the server.

[0520] The server performs image analysis on the received image data using an AI model to analyze the image content in detail. This analysis process utilizes functions such as object recognition and scene detection. Based on the analysis results, the server generates a caption describing the image content.

[0521] In addition, the server uses data from the terminal and an emotion engine to recognize the user's emotions based on their voice responses and behavioral data. This recognized emotion information is used to adjust the tone and content of the generated captions. For example, if the image is a landscape photograph and the user is recognized as being in a relaxed state, the caption may be expressed in a more emotional tone.

[0522] The generated captions are sent to the device, converted into speech using speech synthesis technology, and then provided to the user. Users can listen to the captions as audio, and because appropriate content adjustments are made based on emotion, they can enjoy a richer informational experience.

[0523] In this way, systems equipped with emotion recognition capabilities go beyond simply providing information; they enable flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0524] The following describes the processing flow.

[0525] Step 1:

[0526] A user uses their device to access a website or social networking service and opens a page. During this process, the user performs the usual actions.

[0527] Step 2:

[0528] The device detects images displayed on the open page and retrieves the image data. The images are stored in digital format.

[0529] Step 3:

[0530] The device sends the acquired image data to the server. This transmission is usually done via an API request.

[0531] Step 4:

[0532] The server inputs the received image data into an AI model, which then analyzes the image content. This analysis includes object recognition and scene identification.

[0533] Step 5:

[0534] The server generates image captions based on the analysis results. These captions are written in natural language.

[0535] Step 6:

[0536] The server uses an emotion engine based on audio and behavioral data received from the terminal to recognize the user's emotions. This emotion data is then used to generate captions.

[0537] Step 7:

[0538] The server adjusts the caption's wording, customizing its tone and content based on the recognized emotion.

[0539] Step 8:

[0540] The adjusted caption is sent back to the terminal. The terminal receives the caption and converts it into speech through a speech synthesis engine.

[0541] Step 9:

[0542] Users understand the content of images by listening to audio captions from their devices. Emotion-based tone adjustments provide a more comfortable information experience.

[0543] (Example 2)

[0544] Next, we will describe Example 2. 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."

[0545] Technologies for providing easily understandable image content on the internet to visually impaired individuals are limited, and there is a particular need to provide information that considers the user's emotions in addition to understanding the images. However, conventional systems simply convert image content into audio, which is insufficient in providing information that meets the user's emotional needs. As a result, recipients of information do not have a meaningful experience based on their emotions, which is a problem.

[0546] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0547] In this invention, the server includes means for analyzing image data using AI technology to identify objects and detect scenes, means for generating captions based on the analysis results using generative AI technology, and means for adjusting captions considering the user's emotional state. This enables detailed interpretation of images as well as the provision of information tailored to the user's emotions, thereby providing an emotionally valuable information experience for visually impaired individuals.

[0548] A "terminal" is a device used by users to access image content on the internet and social networks and to collect that data.

[0549] A "server" is a central system that receives image data transmitted from terminals, performs analysis, and processes data based on the results.

[0550] "Image data" refers to a collection of digital data that represents visual information existing on the internet or social networks.

[0551] "AI technology" refers to advanced information processing technology used to analyze image data and perform tasks such as object identification and scene detection.

[0552] "Generative AI technology" is a technology for generating natural language captions from analyzed image data.

[0553] A "caption" is a short sentence generated to describe the content of an image in language.

[0554] "User emotional state" refers to information that indicates the psychological reaction and circumstances of a user when they receive an image.

[0555] "Speech synthesis technology" is a technology that converts text data into speech data, and is a means of providing users with auditory information.

[0556] This invention provides a system for visually impaired individuals to understand image content on the internet and social networks. This system enables flexible information delivery, including a caption generation function that takes user emotions into consideration.

[0557] The user first uses their device to browse websites and social networking services through an internet browser or dedicated application. The device automatically detects and collects image data provided on these web pages. The collected image data is sent to the server using a secure and reliable communication protocol (e.g., HTTPS).

[0558] The server is located on hardware designed to analyze received image data and utilizes high-performance deep learning frameworks (e.g., TensorFlow, PyTorch) as its AI technology. The server leverages techniques such as Convolutional Neural Networks (CNNs) to perform object identification and scene detection, enabling a detailed understanding of the image content.

[0559] Based on the analyzed results, the server generates captions using generative AI technology (e.g., natural language processing models). Prompts are used in this process to instruct the model, organizing the information and generating clear sentences. An example of a prompt might be, "Describe in detail what is depicted in this image."

[0560] Furthermore, the server uses the user's voice data and behavioral logs collected from the terminal to estimate the user's emotional state using an emotion engine. Through speech recognition technology (e.g., speech-to-text services), it analyzes the user's dialogue and tone to evaluate their emotions. Based on the recognized emotion data, it adjusts the tone and content of the generated captions to enable expressions that better reflect the user's emotions.

[0561] Finally, the adjusted captions are sent to the device and converted into speech using speech synthesis technology (e.g., a speech conversion service). By listening to this, users can properly understand the visual information and enjoy a deeper informational experience.

[0562] This system goes beyond simply providing information; it enables flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0563] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0564] Step 1:

[0565] Users browse the internet and social networking pages through their devices. The devices automatically detect and collect image data displayed on these pages. Input includes web page URLs and HTML data via browsers and applications, and output is image data. The devices temporarily store the image data to prepare it for subsequent processing.

[0566] Step 2:

[0567] The terminal sends the collected image data to the server. The input is the image data collected in step 1, and the output is the delivery of the image data to the server. The HTTPS protocol is used for transmission to securely transfer the data. This process prepares the image data for analysis.

[0568] Step 3:

[0569] The server receives transmitted image data and performs analysis using AI technology. The input is image data sent from the terminal, and the output is the analysis results from object recognition and scene detection. As for AI technology, a deep learning framework (e.g., TensorFlow or PyTorch) is used, and CNNs are utilized to identify objects and scenes in the image. This process allows for a detailed understanding of the image content.

[0570] Step 4:

[0571] The server uses the analysis results to generate captions based on generative AI technology. The input is the analysis results obtained in step 3, and the output is a caption that describes the object or scene. The generative AI model uses natural language processing technology, and appropriate captions are generated by providing prompts. For example, a prompt such as "Describe what is in this image" will produce a detailed caption of the image content.

[0572] Step 5:

[0573] The server analyzes user voice data and behavior logs provided by the terminal and evaluates the user's emotions using an emotion engine. The input is voice data and operation logs, and the output is the evaluation result of the user's emotional state. Emotions are inferred from voice tone and operation tendencies based on speech recognition technology and behavior pattern analysis.

[0574] Step 6:

[0575] The server adjusts the caption based on the recognized user's emotions. The input is the caption generated in step 4 and the emotion data evaluated in step 5, and the output is the adjusted caption that matches the emotion. The caption may be changed, for example, to a relaxed tone, to provide information that matches the user's emotional state.

[0576] Step 7:

[0577] The server sends the adjusted caption to the terminal. The input is the adjusted caption obtained in step 6, and the output is the transfer of the data containing this caption to the terminal.

[0578] Step 8:

[0579] The device converts received captions into speech using speech synthesis technology. The input is the caption sent from the server, and the output is audio data that the user can listen to. Speech synthesis technology converts text into natural-sounding speech, which is then provided to the user. As a result, users can understand image information on the internet aurally and experience emotionally adapted content.

[0580] (Application Example 2)

[0581] Next, we will explain application example 2. In the following explanation, 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."

[0582] Visually impaired individuals have difficulty effectively understanding video content on the internet. To address this problem, it is necessary to go beyond simply providing information and adjust the content according to the user's emotional state. While conventional technologies can generate captions that substitute for information the user receives visually, a challenge remains in the lack of a mechanism to dynamically adjust the content in response to the user's emotions.

[0583] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0584] In this invention, the server includes means for receiving and analyzing video data from a terminal, means for analyzing the user's voice response and recognizing emotions, and means for dynamically adjusting the content of caption generation based on the recognized emotions. This allows visually impaired individuals to obtain information appropriate to their emotional state, enabling a more empathetic understanding of the content.

[0585] A "terminal" is an information processing device used by a user, and is a device that has the function of collecting image and audio data and running applications.

[0586] "Motion image data" refers to visual data composed of a series of frames, and includes information in the form of videos and animations.

[0587] A "server" is a centralized computer that receives data from terminals, processes and analyzes it, and provides services to other devices on a communication network.

[0588] "Analysis" is the process of identifying specific patterns or features in received data and deciphering its content.

[0589] A "caption" is a descriptive text generated based on video and image data, representing visual information as textual information.

[0590] "Speech synthesis" is a technology that converts text data into speech, and is a process in which a computer outputs it as spoken language.

[0591] "Voice response" refers to the tone of voice and words spoken by the user through the device, and is information that indicates the user's emotions and state of mind.

[0592] "Emotion recognition" is the process of inferring and identifying a user's emotional state from their voice responses and other data.

[0593] "Adjustment" is the process of modifying the generated content or its presentation according to specific conditions or circumstances.

[0594] This invention provides a system for visually impaired individuals to effectively understand video content and enables flexible caption generation that responds to the user's emotional state. Specific embodiments are described below.

[0595] Suitable devices for users include smartphones and head-mounted displays, which have built-in cameras and microphones. These devices access content via the internet and send video and image data of interest to a server. This transmission is done through an application programming interface.

[0596] The server uses generative AI models for image analysis (e.g., YOLO or ResNet) to analyze the received video data. During this process, object recognition and scene detection are performed. Based on the analysis results, captions describing the content of the video are generated. Furthermore, the server analyzes the user's voice responses from the terminal and uses an emotion recognition engine (e.g., OpenAI's GPT-3) to identify the user's emotional state. This emotion information is used to adjust the content and tone of the generated captions.

[0597] The generated captions are converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and provided to the user. This allows users to complement visual information through emotionally appropriate audio content.

[0598] For example, if a user is watching an action scene in a movie, the emotion recognition results will generate a caption such as "An exciting scene is unfolding." On the other hand, for a quiet, moving scene, the caption will be adjusted to "Heartwarming music is playing." An example of this prompt would be, "Please describe the content of this video scene as an audio guide. Generate a caption while describing the content, using a calm tone if the user is calm, and emphasizing if they are excited."

[0599] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0600] Step 1:

[0601] The terminal retrieves video and image data of interest to the user via the internet and sends it to the server via an application programming interface. The input is video and image data, and the output is data transmission to the server. In this process, the terminal appropriately formats the video and image data and transmits it as packets over the network.

[0602] Step 2:

[0603] The server acquires video data received from the terminal and begins analyzing the data using a generating AI model. The input is video data received from the terminal, and the output is information on object recognition and scene detection as a result of the analysis. The server uses an AI model (e.g., YOLO or ResNet) to identify features of objects and scenes in the video and records this information in a database.

[0604] Step 3:

[0605] The server generates a caption describing the content of the video based on the analysis results. The input is the analysis results from step 2, and the output is the caption text. Natural language generation technology using prompt sentences is applied to generate the caption, expressing the analyzed information in language that is easy for the user to understand.

[0606] Step 4:

[0607] The device collects the user's voice responses and sends them to the server. The input is the user's voice, and the output is the transmission of voice data to the server. The device uses a microphone to capture the voice in real time, compresses the data as needed, and sends it to the server.

[0608] Step 5:

[0609] The server uses a speech recognition engine to analyze the user's emotions from their voice data. The input is the voice data from step 4, and the output is the analyzed user emotion information. The server analyzes the voice waveform and estimates the user's emotional state using template matching and machine learning.

[0610] Step 6:

[0611] The server uses the analyzed sentiment information to adjust the tone and content of the caption. The input is the caption text from step 3 and the sentiment information from step 5, and the output is the adjusted caption. Depending on the sentiment recognition results, the intonation and expression of the caption change, and the final audio output is generated.

[0612] Step 7:

[0613] Finally, the captions generated by the server are converted into speech using speech synthesis technology and sent to the terminal. The input is the adjusted caption text, and the output is audio data. The terminal plays this audio and provides it to the user. Existing technologies such as Google Cloud Text-to-Speech are used for speech synthesis to reproduce natural-sounding speech.

[0614] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0615] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0616] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0617] [Fourth Embodiment]

[0618] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0619] As shown in Figure 7, the 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.

[0620] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0621] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0622] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0623] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0624] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0625] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive 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 robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0626] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0627] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0628] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0629] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0630] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0631] This invention provides a system that enables visually impaired individuals to understand image content on the internet and social networking services. The system includes a terminal, a server, and software with caption generation and speech conversion capabilities.

[0632] Users typically use their devices to navigate websites and social media, opening pages of interest. During this process, the device has a function to automatically detect images on the page. The detected image data is then transmitted to a server via the internet.

[0633] The server uses an AI model to analyze the received image data. This AI model has the ability to analyze the image content in detail using technologies such as object recognition and scene detection. Once the analysis is complete, the server generates text captions corresponding to the images.

[0634] The generated captions are sent back to the device and converted into audio. This conversion uses natural-sounding speech synthesis technology, allowing users to hear the content of the image in audio form.

[0635] For example, if a user opens a travel blog page and tries to view a photograph of a mountain landscape, the image is sent to the server, where a caption such as "The Alps and lush green valleys stretching out under a blue sky" is generated. The device then provides this caption to the user via audio, allowing them to understand the image's content without relying on visual information.

[0636] This system provides a barrier-free environment for visually impaired individuals, enabling them to access images without alternative text, and significantly improving their online information access.

[0637] The following describes the processing flow.

[0638] Step 1:

[0639] A user uses their device to access websites and social networking services and opens pages of interest. During this process, the user performs typical browsing actions.

[0640] Step 2:

[0641] The device automatically detects images on the opened page and temporarily retrieves the image data. The images are stored in URL or base64 format.

[0642] Step 3:

[0643] The terminal prepares to send the acquired image data to the server and sends the image data to the server via an internet connection in the form of an API request.

[0644] Step 4:

[0645] The server receives image data sent from the terminal and begins analyzing the image using an AI model. This analysis includes object recognition and scene detection.

[0646] Step 5:

[0647] The server generates image captions based on the analysis results. The generated captions describe the details of the objects and scenes in the image using natural language.

[0648] Step 6:

[0649] The server sends the generated caption back to the terminal. At this time, the caption is sent in a standard data format.

[0650] Step 7:

[0651] The terminal receives captions from the server and uses a text-to-speech engine to convert the text captions into speech.

[0652] Step 8:

[0653] The device provides the user with an audio caption via a speaker or headphones, allowing the user to understand the image content audibly.

[0654] (Example 1)

[0655] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0656] Individuals with visual impairments face difficulties in effectively understanding visual content on the internet and digital media. Furthermore, systems providing audio information as an alternative to visual information are not sufficiently developed, resulting in a lack of means to accurately convey the meaning of visual content. This situation leads to the challenge of limited information access for individuals with visual impairments.

[0657] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0658] In this invention, the server includes means for analyzing visual information and generating natural language descriptions, means for a mechanism for analyzing visual signals, and means for using sound generation technology to generate naturally sounding synthesized speech. This enables individuals with visual impairments to accurately understand visual information as sound and to smoothly access information on the internet and digital media.

[0659] A "terminal" is a computer device used to detect visual information and transmit it to an information processing device.

[0660] "Visual information" refers to visual data detected from a device as digital images or videos.

[0661] An "information processing device" is a computer device that analyzes visual information received from a terminal and generates natural language descriptions.

[0662] "Natural language description" refers to a form of text that is understandable to humans and is generated based on visual information.

[0663] "Visual signals" refer to image and video data that a device detects and analyzes.

[0664] "Sound generation technology" is a technology that converts text data into speech that sounds like human speech.

[0665] A "learning model" is a computational model that analyzes visual information based on AI technology to perform feature recognition and scene detection.

[0666] This invention provides a system for individuals with visual impairments to understand visual information on the internet and digital media as audio. The system includes a terminal, an information processing device, and software employing sound generation technology to convert visual information into audio.

[0667] Users access web pages and social media content containing visual information through their devices. These devices have the capability to detect visual information on the page and transmit this information to an information processing device. The devices used are commonly available personal computers and smart devices.

[0668] The server, as an information processing unit, is equipped with a learning model for analyzing received visual information. This model is implemented using widely used AI frameworks such as TensorFlow and PyTorch, and recognizes specific features from visual information to generate natural language descriptions. For example, it performs object recognition and scene detection, and then expresses the visual information as text based on the results.

[0669] The generated natural language description is then sent back to the device. The device uses audio generation technologies, such as the Google Text-to-Speech API, to convert the natural language description into synthesized speech. This allows the user to understand visual information through speech.

[0670] As a concrete example, consider a case where a user views a blog post containing a photograph of a tourist destination. The device sends this photograph to the server, which generates a natural language description from the visual information, such as "a historical castle and its surrounding landscape against a blue sky." The device then provides this description to the user as synthesized speech. An example of a prompt could be the instruction, "Please create a caption that describes the content of this image in detail."

[0671] By implementing this invention, individuals with visual impairments can receive visual information as audio, significantly improving their access to information provided online.

[0672] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0673] Step 1:

[0674] The user opens internet or social media pages on their device. The device automatically detects images contained in the page. Specifically, it analyzes the HTML and CSS, This tool extracts data set as background images using tags and CSS. The input is the HTML data of a webpage, and the output is the image URL or binary data.

[0675] Step 2:

[0676] The device sends the detected image data to the server. Specifically, it sends the image data using an HTTP POST request. In this process, the image metadata may also be sent. The input is image data (e.g., JPEG or PNG format), and the output is an HTTP request that is sent to the server.

[0677] Step 3:

[0678] The server inputs the received image data into an AI model. The AI ​​model is, for example, a deep learning model using TensorFlow, and performs object recognition and scene detection. The input is image data, and the output is intermediate data describing the image's features. Specifically, a convolutional neural network (CNN) extracts the main features of the image.

[0679] Step 4:

[0680] The server generates natural language descriptions based on the output of the AI ​​model. The generative AI model analyzes feature data and generates corresponding text. For example, it might create a caption such as "A historical castle spread out under a blue sky." The input is the feature data of the AI ​​model, and the output is a natural language description.

[0681] Step 5:

[0682] The server sends the generated natural language description to the terminal. It returns an HTTP response, transmitting the natural language description to the terminal. The input is the generated natural language description, and the output is the text data sent to the terminal.

[0683] Step 6:

[0684] The device inputs the received natural language description into a speech synthesis engine and generates speech. For example, it uses the Google Text-to-Speech API to convert text into speech. The input is natural language text, and the output is synthesized speech data. Specifically, it converts the text into phonemes and generates a sound waveform.

[0685] Step 7:

[0686] The user listens to audio provided by the device. This allows the user to perceive visual information audibly. The device plays the audio through speakers or headphones. This process includes playback controls and volume adjustments. The input is synthesized audio data, and the output is the audio played back to the user.

[0687] (Application Example 1)

[0688] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0689] Visually impaired individuals face the challenge of understanding online video and image content in real time. Furthermore, general visual caption generation systems are specialized for static images only and are insufficient to assist in understanding dynamic content. Specifically, there is a need for methods to grasp the actions of characters and details of scenes in video works such as movies and television programs in real time, without relying on sight.

[0690] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0691] In this invention, the server includes means for collecting visual information from a terminal and transmitting it to a central processing unit, means for the central processing unit to analyze the visual information and generate a text description, and means for transmitting the generated text description to the terminal and outputting it as audio. This makes it possible for visually impaired people to understand visual information in real time and to enrich their experience of dynamic content.

[0692] A "terminal" is a device that collects visual information and transmits it to a central processing unit.

[0693] "Visual information" refers to visual content such as images and videos, which are analyzed by a central processing unit.

[0694] A "central processing unit" is a system that analyzes visual information transmitted from terminals and generates textual descriptions.

[0695] A "text description" is a written representation of the content of the analyzed visual information, generated to help visually impaired individuals understand its content.

[0696] A "generative model" refers to artificial intelligence technology used to generate appropriate text descriptions from visual information.

[0697] "Audio output" refers to the process of providing the generated text explanation to the user as audio.

[0698] "Real-time" refers to a state where events are processed simultaneously, meaning that information is provided without delay.

[0699] "Dynamic content" refers to content that includes continuously changing visual information, such as videos and live streaming.

[0700] This invention functions as a real-time visual information assistance system for content distribution services. Users use this system on a smart device to view streaming content. The terminal continuously collects visual information and transmits it to a central processing unit for analysis. The central processing unit utilizes artificial intelligence models to analyze images and videos. Its main software technologies include video segmentation using OpenCV and analysis using deep learning models with PyTorch.

[0701] The analyzed data is transformed into a textual description using a generative AI model. This textual description is generated in real time in sync with the streaming and sent back to the device. Finally, the device outputs this textual description as speech using speech synthesis technology, such as the Google Text-to-Speech API, conveying the content to the user in real time.

[0702] As a concrete example, the system analyzes in real time a scene in a movie where a bus suddenly stops in front of a character while the user is watching. An example of a prompt for the generative model would be: "Analyze the movie scene and describe the main characters and their actions in detail. Example: 'A woman in a red dress is standing at the bus stop.' Cover all important actions." This allows visually impaired users to perceive the content of the movie as if they were actually seeing the scene.

[0703] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0704] Step 1:

[0705] The device collects visual information from streaming video data by dividing it into frames. Specifically, it uses software such as OpenCV to capture 30 frames per second and temporarily stores this visual information. The input is streaming video data, and the output is image information divided into frames.

[0706] Step 2:

[0707] The terminal sends the collected visual information as a data request to the central processing unit. Here, each frame is processed as an API request and sent to the server. The input is segmented image information, and the output is the data request sent via the API.

[0708] Step 3:

[0709] The server analyzes the received visual information using an artificial intelligence model. Specifically, it uses a PyTorch-based object recognition model to detect objects and people within each frame. The input is the transmitted frame image, and the output is the detected object information.

[0710] Step 4:

[0711] The server uses a generative AI model to generate text descriptions based on the analysis results. Here, a model such as GPT-3.5 is used to create captions in natural language based on the input object information. The input is object information, and the output is the generated text description.

[0712] Step 5:

[0713] The server sends the generated text description to the terminal. Specifically, it returns text data to the terminal as a data response. The input is the generated text description, and the output is the text data sent to the terminal.

[0714] Step 6:

[0715] The device outputs the received text description as audio. Specifically, it uses the Google Text-to-Speech API to play the caption as synthesized speech. The input is the text description received from the server, and the output is audio data that the user can listen to.

[0716] Through the processing steps described above, visually impaired individuals can grasp visual information in real time and obtain details of dynamic content via audio.

[0717] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0718] This invention provides a system for visually impaired individuals to understand image content via the internet and social networking services, and also incorporates an emotion engine for generating captions that take user emotions into account. This system includes a terminal, a server, and a speech-generating function, as well as an emotion engine with emotion recognition capabilities.

[0719] Users use their devices to browse websites and social media, accessing pages that interest them. The devices automatically collect image data from these pages and send it to the server.

[0720] The server performs image analysis on the received image data using an AI model to analyze the image content in detail. This analysis process utilizes functions such as object recognition and scene detection. Based on the analysis results, the server generates a caption describing the image content.

[0721] In addition, the server uses data from the terminal and an emotion engine to recognize the user's emotions based on their voice responses and behavioral data. This recognized emotion information is used to adjust the tone and content of the generated captions. For example, if the image is a landscape photograph and the user is recognized as being in a relaxed state, the caption may be expressed in a more emotional tone.

[0722] The generated captions are sent to the device, converted into speech using speech synthesis technology, and then provided to the user. Users can listen to the captions as audio, and because appropriate content adjustments are made based on emotion, they can enjoy a richer informational experience.

[0723] In this way, systems equipped with emotion recognition capabilities go beyond simply providing information; they enable flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0724] The following describes the processing flow.

[0725] Step 1:

[0726] A user uses their device to access a website or social networking service and opens a page. During this process, the user performs the usual actions.

[0727] Step 2:

[0728] The device detects images displayed on the open page and retrieves the image data. The images are stored in digital format.

[0729] Step 3:

[0730] The device sends the acquired image data to the server. This transmission is usually done via an API request.

[0731] Step 4:

[0732] The server inputs the received image data into an AI model and analyzes the image content. This analysis includes object recognition and scene identification.

[0733] Step 5:

[0734] The server generates image captions based on the analysis results. These captions are written in natural language.

[0735] Step 6:

[0736] The server uses an emotion engine based on audio and behavioral data received from the terminal to recognize the user's emotions. This emotion data is then used to generate captions.

[0737] Step 7:

[0738] The server adjusts the caption's wording, customizing its tone and content based on the recognized emotion.

[0739] Step 8:

[0740] The adjusted caption is sent back to the terminal. The terminal receives the caption and converts it into speech through a speech synthesis engine.

[0741] Step 9:

[0742] Users understand the content of images by listening to audio captions from their devices. Emotion-based tone adjustments provide a more comfortable information experience.

[0743] (Example 2)

[0744] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0745] Technologies for providing easily understandable image content on the internet to visually impaired individuals are limited, and there is a particular need to provide information that considers the user's emotions in addition to understanding the images. However, conventional systems simply convert image content into audio, which is insufficient in providing information that meets the user's emotional needs. As a result, recipients of information do not have a meaningful experience based on their emotions, which is a problem.

[0746] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0747] In this invention, the server includes means for analyzing image data using AI technology to identify objects and detect scenes, means for generating captions based on the analysis results using generative AI technology, and means for adjusting captions considering the user's emotional state. This enables detailed interpretation of images as well as the provision of information tailored to the user's emotions, thereby providing an emotionally valuable information experience for visually impaired individuals.

[0748] A "terminal" is a device used by users to access image content on the internet and social networks and to collect that data.

[0749] A "server" is a central system that receives image data transmitted from terminals, performs analysis, and processes data based on the results.

[0750] "Image data" refers to a collection of digital data that represents visual information existing on the internet or social networks.

[0751] "AI technology" refers to advanced information processing technology used to analyze image data and perform tasks such as object identification and scene detection.

[0752] "Generative AI technology" is a technology for generating natural language captions from analyzed image data.

[0753] A "caption" is a short sentence generated to describe the content of an image in language.

[0754] "User emotional state" refers to information that indicates the psychological reaction and circumstances of a user when they receive an image.

[0755] "Speech synthesis technology" is a technology that converts text data into speech data, and is a means of providing users with auditory information.

[0756] This invention provides a system for visually impaired individuals to understand image content on the internet and social networks. This system enables flexible information delivery, including a caption generation function that takes user emotions into consideration.

[0757] The user first uses their device to browse websites and social networking services through an internet browser or dedicated application. The device automatically detects and collects image data provided on these web pages. The collected image data is sent to the server using a secure and reliable communication protocol (e.g., HTTPS).

[0758] The server is located on hardware designed to analyze received image data and utilizes high-performance deep learning frameworks (e.g., TensorFlow, PyTorch) as its AI technology. The server leverages techniques such as Convolutional Neural Networks (CNNs) to perform object identification and scene detection, enabling a detailed understanding of the image content.

[0759] Based on the analyzed results, the server generates captions using generative AI technology (e.g., natural language processing models). Prompts are used in this process to instruct the model, organizing the information and generating clear sentences. An example of a prompt might be, "Describe in detail what is depicted in this image."

[0760] Furthermore, the server uses the user's voice data and behavioral logs collected from the terminal to estimate the user's emotional state using an emotion engine. Through speech recognition technology (e.g., speech-to-text services), it analyzes the user's dialogue and tone to evaluate their emotions. Based on the recognized emotion data, it adjusts the tone and content of the generated captions to enable expressions that better reflect the user's emotions.

[0761] Finally, the adjusted captions are sent to the device and converted into speech using speech synthesis technology (e.g., a speech conversion service). By listening to this, users can properly understand the visual information and enjoy a deeper informational experience.

[0762] This system goes beyond simply providing information; it enables flexible content generation that responds to users' emotional needs, further improving the quality of information access for visually impaired individuals.

[0763] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0764] Step 1:

[0765] Users browse the internet and social networking pages through their devices. The devices automatically detect and collect image data displayed on these pages. Input includes web page URLs and HTML data via browsers and applications, and output is image data. The devices temporarily store the image data to prepare it for subsequent processing.

[0766] Step 2:

[0767] The terminal sends the collected image data to the server. The input is the image data collected in step 1, and the output is the delivery of the image data to the server. The HTTPS protocol is used for transmission to securely transfer the data. This process prepares the image data for analysis.

[0768] Step 3:

[0769] The server receives transmitted image data and performs analysis using AI technology. The input is image data sent from the terminal, and the output is the analysis results from object recognition and scene detection. As for AI technology, a deep learning framework (e.g., TensorFlow or PyTorch) is used, and CNNs are utilized to identify objects and scenes in the image. This process allows for a detailed understanding of the image content.

[0770] Step 4:

[0771] The server uses the analysis results to generate captions based on generative AI technology. The input is the analysis results obtained in step 3, and the output is a caption that describes the object or scene. The generative AI model uses natural language processing technology, and appropriate captions are generated by providing prompts. For example, a prompt such as "Describe what is in this image" will produce a detailed caption of the image content.

[0772] Step 5:

[0773] The server analyzes user voice data and behavior logs provided by the terminal and evaluates the user's emotions using an emotion engine. The input is voice data and operation logs, and the output is the evaluation result of the user's emotional state. Emotions are inferred from voice tone and operation tendencies based on speech recognition technology and behavior pattern analysis.

[0774] Step 6:

[0775] The server adjusts the caption based on the recognized user's emotions. The input is the caption generated in step 4 and the emotion data evaluated in step 5, and the output is the adjusted caption that matches the emotion. The caption may be changed, for example, to a relaxed tone, to provide information that matches the user's emotional state.

[0776] Step 7:

[0777] The server sends the adjusted caption to the terminal. The input is the adjusted caption obtained in step 6, and the output is the transfer of the data containing this caption to the terminal.

[0778] Step 8:

[0779] The device converts received captions into speech using speech synthesis technology. The input is the caption sent from the server, and the output is audio data that the user can listen to. Speech synthesis technology converts text into natural-sounding speech, which is then provided to the user. As a result, users can understand image information on the internet aurally and experience emotionally adapted content.

[0780] (Application Example 2)

[0781] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0782] Visually impaired individuals have difficulty effectively understanding video content on the internet. To address this problem, it is necessary to go beyond simply providing information and adjust the content according to the user's emotional state. While conventional technologies can generate captions that substitute for information the user receives visually, a challenge remains in the lack of a mechanism to dynamically adjust the content in response to the user's emotions.

[0783] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0784] In this invention, the server includes means for receiving and analyzing video data from a terminal, means for analyzing the user's voice response and recognizing emotions, and means for dynamically adjusting the content of caption generation based on the recognized emotions. This allows visually impaired individuals to obtain information appropriate to their emotional state, enabling a more empathetic understanding of the content.

[0785] A "terminal" is an information processing device used by a user, and is a device that has the function of collecting image and audio data and running applications.

[0786] "Motion image data" refers to visual data composed of a series of frames, and includes information in the form of videos and animations.

[0787] A "server" is a centralized computer that receives data from terminals, processes and analyzes it, and provides services to other devices on a communication network.

[0788] "Analysis" is the process of identifying specific patterns or features in received data and deciphering its content.

[0789] A "caption" is a descriptive text generated based on video and image data, representing visual information as textual information.

[0790] "Speech synthesis" is a technology that converts text data into speech, and is a process in which a computer outputs it as spoken language.

[0791] "Voice response" refers to the tone of voice and words spoken by the user through the device, and is information that indicates the user's emotions and state of mind.

[0792] "Emotion recognition" is the process of inferring and identifying a user's emotional state from their voice responses and other data.

[0793] "Adjustment" is the process of modifying the generated content or its presentation according to specific conditions or circumstances.

[0794] This invention provides a system for visually impaired individuals to effectively understand video content and enables flexible caption generation that responds to the user's emotional state. Specific embodiments are described below.

[0795] Suitable devices for users include smartphones and head-mounted displays, which have built-in cameras and microphones. These devices access content via the internet and send video and image data of interest to a server. This transmission is done through an application programming interface.

[0796] The server uses generative AI models for image analysis (e.g., YOLO or ResNet) to analyze the received video data. During this process, object recognition and scene detection are performed. Based on the analysis results, captions describing the content of the video are generated. Furthermore, the server analyzes the user's voice responses from the terminal and uses an emotion recognition engine (e.g., OpenAI's GPT-3) to identify the user's emotional state. This emotion information is used to adjust the content and tone of the generated captions.

[0797] The generated captions are converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and provided to the user. This allows users to complement visual information through emotionally appropriate audio content.

[0798] For example, if a user is watching an action scene in a movie, the emotion recognition results will generate a caption such as "An exciting scene is unfolding." On the other hand, for a quiet, moving scene, the caption will be adjusted to "Heartwarming music is playing." An example of this prompt would be, "Please describe the content of this video scene as an audio guide. Generate a caption while describing the content, using a calm tone if the user is calm, and emphasizing if they are excited."

[0799] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0800] Step 1:

[0801] The terminal retrieves video and image data of interest to the user via the internet and sends it to the server via an application programming interface. The input is video and image data, and the output is data transmission to the server. In this process, the terminal appropriately formats the video and image data and transmits it as packets over the network.

[0802] Step 2:

[0803] The server acquires video data received from the terminal and begins analyzing the data using a generating AI model. The input is video data received from the terminal, and the output is information on object recognition and scene detection as a result of the analysis. The server uses an AI model (e.g., YOLO or ResNet) to identify features of objects and scenes in the video and records this information in a database.

[0804] Step 3:

[0805] The server generates a caption describing the content of the video based on the analysis results. The input is the analysis results from step 2, and the output is the caption text. Natural language generation technology using prompt sentences is applied to generate the caption, expressing the analyzed information in language that is easy for the user to understand.

[0806] Step 4:

[0807] The device collects the user's voice responses and sends them to the server. The input is the user's voice, and the output is the transmission of voice data to the server. The device uses a microphone to capture the voice in real time, compresses the data as needed, and sends it to the server.

[0808] Step 5:

[0809] The server uses a speech recognition engine to analyze the user's emotions from their voice data. The input is the voice data from step 4, and the output is the analyzed user emotion information. The server analyzes the voice waveform and estimates the user's emotional state using template matching and machine learning.

[0810] Step 6:

[0811] The server uses the analyzed sentiment information to adjust the tone and content of the caption. The input is the caption text from step 3 and the sentiment information from step 5, and the output is the adjusted caption. Depending on the sentiment recognition results, the intonation and expression of the caption change, and the final audio output is generated.

[0812] Step 7:

[0813] Finally, the captions generated by the server are converted into speech using speech synthesis technology and sent to the terminal. The input is the adjusted caption text, and the output is audio data. The terminal plays this audio and provides it to the user. Existing technologies such as Google Cloud Text-to-Speech are used for speech synthesis to reproduce natural-sounding speech.

[0814] 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 controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0815] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0816] 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 this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0817] Furthermore, the emotion identification model 59, acting 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 a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0818] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0819] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0820] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0821] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0822] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is 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 the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0823] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0824] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0825] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0826] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0827] 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.

[0828] Furthermore, it is not necessary to store the entirety 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 the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0829] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0830] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of 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). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0831] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0832] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0833] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0834] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0835] The following is further disclosed regarding the embodiments described above.

[0836] (Claim 1)

[0837] A means of collecting image data from a terminal and sending it to a server,

[0838] A means for analyzing image data on a server and generating captions,

[0839] A means of sending the generated caption to a terminal and converting it into speech,

[0840] A system that includes this.

[0841] (Claim 2)

[0842] The system according to claim 1, which sends image data collected by a terminal to a server as an API request.

[0843] (Claim 3)

[0844] The system according to claim 1, wherein the server uses an AI model to analyze image data and perform object recognition and scene detection.

[0845]

[0846] "Example 1"

[0847] (Claim 1)

[0848] A means for detecting visual information from a terminal and transmitting it to an information processing device,

[0849] A means for analyzing visual information using an information processing device and generating a natural language description,

[0850] A means of sending the generated natural language description to a terminal and converting it into speech,

[0851] In detecting visual information, a means equipped with a mechanism for analyzing visual signals,

[0852] A means of using sound generation technology to generate naturally sounding synthesized speech,

[0853] A system that includes this.

[0854] (Claim 2)

[0855] The system according to claim 1, which transmits visual information detected by a terminal to an information processing device in a data exchange format.

[0856] (Claim 3)

[0857] The system according to claim 1, wherein the information processing device analyzes visual information using a learning model and performs feature recognition and scene detection.

[0858] "Application Example 1"

[0859] (Claim 1)

[0860] A means for collecting visual information from a terminal and transmitting it to a central processing unit,

[0861] A means for analyzing visual information in a central processing unit and generating textual descriptions,

[0862] A means of sending the generated text description to a terminal and outputting it as audio,

[0863] A means of processing visual information in real time and assisting in the delivery of dynamic content,

[0864] A means of generating text descriptions using a generative model,

[0865] A system that includes this.

[0866] (Claim 2)

[0867] The system according to claim 1, wherein visual information collected by a terminal is transmitted to a central processing unit as a data request.

[0868] (Claim 3)

[0869] The system according to claim 1, wherein the central processing unit analyzes visual information using an artificial intelligence model and performs object recognition, scene detection, and text description generation.

[0870] "Example 2 of combining an emotion engine"

[0871] (Claim 1)

[0872] A means of collecting image data from a terminal and sending it to a server,

[0873] A means for analyzing image data on a server using AI technology to identify objects and detect scenes,

[0874] A means of generating captions using AI generation technology based on analysis results, and adjusting the captions while considering the user's emotional state,

[0875] A means for sending the generated caption to a terminal and converting it into speech using speech synthesis technology,

[0876] A system that includes this.

[0877] (Claim 2)

[0878] The system according to claim 1, wherein image data collected by a terminal is transmitted to a server as a communication means.

[0879] (Claim 3)

[0880] The system according to claim 1, wherein the server uses AI technology to analyze image data, identify objects and detect scenes, evaluate the user's emotions, and use this to adjust the generated captions.

[0881] "Application example 2 when combining with an emotional engine"

[0882] (Claim 1)

[0883] A means of collecting video data from a terminal and sending it to a server,

[0884] A means for analyzing video data on a server and generating captions,

[0885] A means of sending the generated caption to a terminal and converting it into speech using speech synthesis,

[0886] A means of analyzing the user's voice responses and recognizing emotions,

[0887] Means for adjusting the content of captions based on perceived emotions,

[0888] A system that includes this.

[0889] (Claim 2)

[0890] The system according to claim 1, wherein motion image data collected by a terminal is sent to a server as an application programming interface request.

[0891] (Claim 3)

[0892] The system according to claim 1, wherein the server analyzes motion image data using a generative model, performs object recognition and scene detection, and further adjusts the captions taking into account the user's emotions. [Explanation of Symbols]

[0893] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of collecting image data from a terminal and sending it to a server, A means for analyzing image data on a server and generating captions, A means of sending the generated caption to a terminal and converting it into speech, A system that includes this.

2. The system according to claim 1, wherein image data collected by a terminal is sent to a server as an API request.

3. The system according to claim 1, wherein the server uses an AI model to analyze image data and perform object recognition and scene detection.

Images