system

Abstract

The system according to this embodiment aims to construct a high-precision dataset specifically for Japanese culture. [Solution] The system according to the embodiment comprises a base labeling unit, an annotation reception unit, and a dataset construction unit. The base labeling unit performs base labeling of images. The annotation reception unit annotates ambiguous parts of the images labeled by the base labeling unit. The dataset construction unit constructs a dataset specialized for Japanese culture based on the data annotated by the annotation reception unit.

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 in 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] In the conventional technology, there is a possibility of giving a sense of discomfort to Japanese users by a CAPTCHA system centered on overseas images.

[0005] The system according to the embodiment aims to construct a high-precision dataset specialized for Japanese culture.

Means for Solving the Problems

[0006] The system according to this embodiment comprises a base labeling unit, an annotation reception unit, and a dataset construction unit. The base labeling unit performs base labeling of images. The annotation reception unit annotates ambiguous parts of the images labeled by the base labeling unit. The dataset construction unit constructs a dataset specific to Japanese culture based on the data annotated by the annotation reception unit. [Effects of the Invention]

[0007] The system according to this embodiment can construct a high-precision dataset specifically for Japanese culture. [Brief explanation of the drawing]

[0008] [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] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This 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] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This 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] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This 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] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Modes for carrying out the invention]

[0009] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0010] First, let's explain the terminology used in the following explanation.

[0011] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Furthermore, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), or TPU (Tensor Processing Unit).

[0012] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

[0013] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0014] In the following embodiments, the labeled communication I / F (Interface) is an interface including a communication processor, an antenna, and the like. The communication I / F manages communication between a plurality of 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).

[0015] 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 only A, only B, or a combination of A and B. Also, in this specification, when expressing three or more matters connected by "and / or", the same concept as "A and / or B" is applied.

[0016] [First Embodiment] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0017] As shown in FIG. 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.

[0018] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. Also, the database 24 and the communication I / F 26 are connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0019] The smart device 14 comprises a computer 36, a receiving 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 receiving device 38, output device 40, and camera 42 are also connected to the bus 52.

[0020] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, and accepts user input. The touch panel 38A accepts user input via touch by detecting contact with an object (e.g., a pen or finger). The microphone 38B accepts user input via voice 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 unit 12. In the data processing unit 12, the specific processing unit 290 (see Figure 2) acquires the data indicating the user input.

[0021] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user by outputting the data in a form perceptible to the user (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.

[0022] 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.

[0023] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0024] 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.

[0025] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0026] In the smart device 14, specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The specific processing program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 operating as a control unit 46A according to the specific processing program 60 executed on the RAM 48. The smart device 14 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0027] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device having the data generation model 58. The data processing device 12 may also be a server device or a terminal device owned by a user (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of processing by the data processing system 10 according to the first embodiment will be described.

[0028] (Example of form 1) The CAPTCHA system according to an embodiment of the present invention is a system in which Japanese users annotate Japanese culture. This CAPTCHA system solves the problem that conventional CAPTCHA systems mainly use images from overseas, which often felt unnatural to Japanese users. The CAPTCHA system aims to build a high-precision dataset specializing in Japanese culture and cultivate an AI model suitable for the Japanese market by having an AI agent perform base labeling and having Japanese users fill in ambiguous parts through CAPTCHA. For example, in the CAPTCHA system, the AI ​​agent first performs base labeling of images. Next, for ambiguous parts, Japanese users annotate through CAPTCHA. In this process, the user looks at images related to Japanese culture, recognizes their content, and labels them. For example, they are required to assign appropriate labels to images of traditional Japanese festivals, food, landscapes, etc. Through this mechanism, a high-precision dataset specializing in Japanese culture is efficiently collected. Furthermore, the AI ​​agent continuously learns and gains a deeper understanding of the unique characteristics of Japanese culture, enabling highly accurate annotation. As a result, an AI model suitable for the Japanese market is cultivated, and it is expected to have applications in various fields. For example, in the tourism industry, the development of AI models that provide information about Japanese tourist destinations and culture will enable the provision of more appropriate information to tourists. Furthermore, in the education sector, the development of AI models that automatically generate educational materials on Japanese history and culture is expected to improve the quality of education. Thus, by building high-precision datasets specifically for Japanese culture and cultivating AI models suited to the Japanese market, CAPTCHA systems are expected to have applications in various fields. This allows CAPTCHA systems to build high-precision datasets specifically for Japanese culture and cultivate AI models suited to the Japanese market.

[0029] The CAPTCHA system according to this embodiment comprises a base labeling unit, an annotation reception unit, and a dataset construction unit. The base labeling unit performs base labeling of images. The base labeling unit performs base labeling of images using, for example, an AI agent. The AI ​​agent can analyze the content of the image and assign appropriate labels. For example, the AI ​​agent can recognize objects in the image and assign corresponding labels to them. The AI ​​agent can also perform labeling while considering the background information of the image. For example, if the background of the image contains specific cultural elements, the labeling will take those elements into consideration. The annotation reception unit annotates ambiguous parts of the image that has been labeled by the base labeling unit. For example, the annotation reception unit receives annotations from Japanese users via CAPTCHA. The user looks at the image displayed via CAPTCHA, recognizes its content, and assigns appropriate labels. For example, they are required to assign appropriate labels to images of traditional Japanese festivals, food, landscapes, etc. The dataset construction unit constructs a dataset specializing in Japanese culture based on the data annotated by the annotation reception unit. The dataset construction unit, for example, evaluates the annotation results and constructs a high-precision dataset. The dataset construction unit clarifies the evaluation criteria and methods for the annotation results and improves the accuracy of the dataset. As a result, the CAPTCHA system according to this embodiment can construct a high-precision dataset specialized for Japanese culture and cultivate an AI model suitable for the Japanese market.

[0030] The base labeling unit performs base labeling of images. For example, the base labeling unit uses an AI agent to perform base labeling of images. The AI ​​agent can analyze the content of an image and assign appropriate labels. Specifically, the AI ​​agent uses deep learning technology to recognize objects within an image and assigns corresponding labels. For example, it recognizes objects unique to Japan, such as tatami rooms, shrine gates, and people wearing traditional Japanese clothing, and assigns labels such as "tatami," "torii," and "kimono" to them, respectively. The AI ​​agent can also consider the background information of the image when labeling. For example, if cherry blossoms are blooming in the background of the image, it will consider this element and assign labels such as "spring" and "cherry blossoms." Furthermore, in image analysis, the AI ​​agent utilizes object detection algorithms and semantic segmentation techniques to analyze each element within the image in detail. This allows the base labeling unit to analyze the content of images with high accuracy and assign appropriate labels. The AI ​​agent has been pre-trained on a large dataset of Japanese culture data, enabling it to recognize objects and background elements unique to Japan with high accuracy. This enables the base labeling unit to achieve highly accurate labeling specifically for Japanese culture, contributing to improved accuracy in subsequent annotation reception and dataset construction units.

[0031] The annotation reception unit annotates ambiguous parts of images labeled by the base labeling unit. For example, Japanese users perform annotation through CAPTCHA. Users view the image displayed through CAPTCHA, recognize its content, and assign appropriate labels. Specifically, users verify whether the labels assigned by the base labeling unit are correct for the displayed image and add corrections or additional labels as needed. For example, they are required to assign appropriate labels to images of traditional Japanese festivals, food, and landscapes. Users can consider details and background information within the image to assign more specific labels. For example, they might assign detailed labels such as "summer festival," "fireworks," and "yukata" to an image of a festival. The annotation reception unit also improves usability by designing the interface for users to perform annotation. For example, it provides features such as the ability to enlarge images and suggest label candidates to enable users to perform annotation efficiently. Furthermore, the annotation reception unit collects user annotation results in real time and stores them in a database. This allows the annotation reception unit to leverage the user's knowledge and experience to supplement ambiguous areas that the base labeling unit may have missed, thereby achieving highly accurate labeling.

[0032] The Dataset Construction Department builds datasets specifically for Japanese culture based on data annotated by the Annotation Reception Department. For example, the Dataset Construction Department evaluates annotation results to build highly accurate datasets. Specifically, it clarifies evaluation criteria and methods for annotation results to improve dataset accuracy. For instance, to evaluate the consistency and accuracy of annotation results, it compares annotation results from multiple users and selects reliable labels. Furthermore, the Dataset Construction Department introduces algorithms to automatically evaluate annotation results, efficiently building highly accurate datasets. In addition, the Dataset Construction Department regularly updates the constructed datasets to reflect the latest information. For example, it adds newly collected images and annotation results to enrich the dataset content. This ensures that the Dataset Construction Department can consistently provide highly accurate datasets based on the latest information. Finally, the Dataset Construction Department utilizes the constructed datasets to train other AI models, developing AI models suitable for the Japanese market. For example, it uses highly accurate datasets related to traditional Japanese festivals, food, and landscapes to train image recognition models and natural language processing models. This allows the dataset building unit to construct highly accurate datasets specifically for Japanese culture and cultivate AI models suitable for the Japanese market.

[0033] The base labeling unit can perform base labeling of images using an AI agent. For example, the base labeling unit can have the AI ​​agent analyze the content of the image and assign an appropriate label. For example, the AI ​​agent can recognize objects in the image and assign a corresponding label to them. The base labeling unit can also have the AI ​​agent consider the background information of the image when performing labeling. For example, if the background of the image contains specific cultural elements, the unit will take those elements into consideration when labeling. This improves the accuracy of base labeling by using an AI agent. Some or all of the above processing in the base labeling unit may be performed using a generative AI, or it may be performed without a generative AI. For example, the base labeling unit can input image data into a generative AI, analyze the content of the image, and perform labeling.

[0034] The annotation reception unit allows Japanese users to perform annotations via CAPTCHA. For example, the annotation reception unit can view images displayed by Japanese users through CAPTCHA, recognize their content, and assign appropriate labels. For instance, it is required to assign appropriate labels to images of traditional Japanese festivals, food, landscapes, etc. The annotation reception unit can also provide an interface for users to perform annotations. For example, it can provide a user-friendly screen layout and input method. This allows for the construction of a dataset specifically focused on Japanese culture through annotations by Japanese users. Some or all of the above-described processes in the annotation reception unit may be performed using AI, or not. For example, the annotation reception unit can input user input data into a generating AI, analyze the annotation content, and perform labeling.

[0035] The dataset construction unit can construct high-precision datasets specifically for Japanese culture. For example, the dataset construction unit constructs datasets based on data annotated by the annotation reception unit. The dataset construction unit evaluates the annotation results and constructs high-precision datasets. For example, the dataset construction unit clarifies the evaluation criteria and methods for annotation results to improve the accuracy of the dataset. By doing so, it is possible to cultivate AI models suitable for the Japanese market by constructing high-precision datasets specifically for Japanese culture. Some or all of the above-described processes in the dataset construction unit may be performed using AI, for example, or without using AI. For example, the dataset construction unit can input annotation data into a generation AI and have the generation AI execute the dataset construction.

[0036] The base labeling unit can continuously learn and gain a deeper understanding of the unique characteristics of Japanese culture. For example, the base labeling unit can set the update frequency of the training data and periodically learn new data. For example, the base labeling unit can collect new image data related to Japanese culture and add it to the training data. The base labeling unit can also optimize the learning algorithm to improve the efficiency of learning. For example, the base labeling unit can use a deep learning algorithm to more accurately capture image features. As a result, the accuracy of base labeling improves through continuous learning. Some or all of the above processes in the base labeling unit may be performed using AI, for example, or not using AI. For example, the base labeling unit can input training data into a generating AI and have the generating AI perform the optimization of the learning algorithm.

[0037] The annotation reception unit can provide an interface for users to perform annotations. For example, the annotation reception unit can provide a user-friendly screen layout and input method. For instance, it can provide a zoom function to make images easier for users to view and a dropdown menu to make label selection easier. The annotation reception unit can also provide guidelines and help functions for users to perform annotations. For example, it can provide guidelines that users can refer to when they get lost and a help function that explains how to operate the system. This improves the efficiency of annotation by providing an interface that makes it easy for users to perform annotations. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or not using AI. For example, the annotation reception unit can input user operation data into a generating AI and have the generating AI perform interface optimization.

[0038] The dataset construction unit can evaluate annotation results. For example, the dataset construction unit can clarify evaluation criteria and methods for annotation results to improve the accuracy of the dataset. For example, the dataset construction unit can evaluate the reliability of annotation results and prioritize the inclusion of highly reliable data in the dataset. The dataset construction unit can also evaluate the consistency of annotation results and include consistent data in the dataset. For example, the dataset construction unit can compare annotation results from multiple users and select consistent labels. This improves the accuracy of the dataset by evaluating the annotation results. Some or all of the above processes in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input annotation data into a generating AI and have the generating AI set evaluation criteria and perform the evaluation.

[0039] The base labeling unit can improve the accuracy of labeling based on the background information of the image. For example, if the background of the image contains specific cultural elements, the base labeling unit will take those elements into consideration when labeling. For example, if the background of the image is complex, the base labeling unit will have an AI agent analyze the background information to improve the accuracy of labeling. Also, if the background of the image is simple, the base labeling unit can have an AI agent focus on the main object and perform the labeling. This improves the accuracy of labeling by considering the background information of the image. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or without AI. For example, the base labeling unit can input image data into a generating AI, analyze the background information, and perform labeling.

[0040] The base labeling unit can adjust the level of detail in the labeling based on the date and time and location where the image was taken. For example, if the image was taken in a particular season, the base labeling unit will add a label related to that season. For example, if the image was taken in a particular location, the base labeling unit will add a label related to that location. Furthermore, if the date and time the image was taken is important, the base labeling unit can also adjust the level of detail in the labeling based on that date and time. This allows for more accurate labeling by adjusting the level of detail in the labeling based on the date and time and location where the image was taken. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or not using AI. For example, the base labeling unit can input the image metadata into a generating AI and have the generating AI perform the adjustment of the level of detail in the labeling based on the date and time and location where the image was taken.

[0041] The base labeling unit can perform labeling based on the color information of the image. For example, if the colors of the image have a specific cultural meaning, the base labeling unit will take that meaning into consideration when labeling. For example, if the colors of the image are vivid, the base labeling unit will perform labeling based on the dominant color. Also, if the colors of the image are muted, the base labeling unit can perform labeling by considering the overall color tone. This improves the accuracy of labeling by considering the color information of the image. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or without AI. For example, the base labeling unit can input image data into a generating AI, analyze the color information, and perform labeling.

[0042] The base labeling unit can improve the accuracy of labeling by analyzing the composition and elements of an image. For example, if the composition of an image has a specific cultural meaning, the base labeling unit will consider that meaning when labeling. For example, if the elements of an image are complex, the base labeling unit will focus on the main elements when labeling. Also, if the elements of an image are simple, the base labeling unit can consider the overall composition when labeling. In this way, the accuracy of labeling is improved by analyzing the composition and elements of an image. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or without AI. For example, the base labeling unit can input image data into a generating AI, analyze the composition and elements, and perform labeling.

[0043] The annotation reception unit can provide the optimal interface based on the user's past annotation history. For example, the annotation reception unit can provide the optimal interface based on the interface the user has used in the past. For example, the annotation reception unit can propose the most efficient interface from the user's past annotation history. The annotation reception unit can also analyze the user's past annotation history and provide the optimal interface. This improves the efficiency of annotation by providing the optimal interface based on the user's past annotation history. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without AI. For example, the annotation reception unit can input the user's annotation history data into a generating AI and have the generating AI perform the task of providing the optimal interface.

[0044] The annotation reception unit can customize the content of annotations based on the user's expertise and interests. For example, if a user is knowledgeable in a particular field, the annotation reception unit will prioritize providing annotations related to that field. For example, the annotation reception unit will provide relevant annotations based on the user's interests. The annotation reception unit can also provide optimal annotation content by considering the user's expertise. This improves the accuracy of annotations by customizing the content of annotations based on the user's expertise and interests. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without AI. For example, the annotation reception unit can input user profile data into a generating AI and have the generating AI perform the customization of the annotation content.

[0045] The annotation reception unit can prioritize receiving annotations that are highly relevant based on the user's geographical location information. For example, if the user is in a specific region, the annotation reception unit will prioritize providing annotations related to that region. For example, the annotation reception unit will provide the most suitable annotations based on the user's geographical location information. The annotation reception unit can also prioritize receiving annotations that are highly relevant, taking into account the user's geographical location information. This improves the accuracy of annotations by prioritizing the acceptance of annotations that are highly relevant based on the user's geographical location information. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without using AI. For example, the annotation reception unit can input the user's geographical location data into a generating AI and have the generating AI perform the task of providing highly relevant annotations.

[0046] The annotation reception unit can accept relevant annotations based on the user's social media activity. For example, the annotation reception unit can provide annotations related to the user's areas of interest based on their social media activity. For example, the annotation reception unit can provide the most suitable annotations based on the user's social media activity. The annotation reception unit can also analyze the user's social media activity and prioritize accepting relevant annotations. This improves the accuracy of annotations by accepting relevant annotations based on the user's social media activity. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without AI. For example, the annotation reception unit can input the user's social media data into a generating AI and have the generating AI perform the task of providing relevant annotations.

[0047] The dataset construction unit can apply the optimal construction algorithm based on past datasets. For example, the dataset construction unit can select the optimal construction algorithm based on past datasets. For example, the dataset construction unit can apply an efficient construction algorithm from past datasets. The dataset construction unit can also apply the optimal construction algorithm by referring to past datasets. This improves the accuracy of the dataset by applying the optimal construction algorithm based on past datasets. Some or all of the above processes in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input past datasets into a generation AI and have the generation AI select and apply the optimal construction algorithm.

[0048] The dataset construction unit can apply different construction methods to each image category. For example, the dataset construction unit can select the optimal construction method for each image category. For example, the dataset construction unit can apply different construction methods depending on the image category. Alternatively, the dataset construction unit can apply the optimal construction method to each image category. By applying different construction methods to each image category, the accuracy of the dataset is improved. Some or all of the above processing in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input image data into a generation AI and have the generation AI execute the application of different construction methods for each category.

[0049] The dataset construction unit can adjust the dataset structure based on the date, time, and location where the images were taken. For example, if an image was taken in a particular season, the dataset construction unit can construct a dataset related to that season. For example, if an image was taken in a particular location, the dataset construction unit can construct a dataset related to that location. Furthermore, if the date and time the images were taken is important, the dataset construction unit can also adjust the dataset structure based on that date and time. This allows for the construction of a more accurate dataset by adjusting the dataset structure based on the date, time, and location where the images were taken. Some or all of the above processing in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input image metadata into a generating AI and have the generating AI perform adjustments to the dataset structure based on the date, time, and location where the images were taken.

[0050] The dataset construction unit can improve the accuracy of the dataset based on the relevant literature for the images. For example, the dataset construction unit improves the accuracy of the dataset based on the relevant literature for the images. For example, the dataset construction unit constructs an optimal dataset by referring to the relevant literature for the images. The dataset construction unit can also improve the accuracy of the dataset based on the relevant literature for the images. By improving the accuracy of the dataset based on the relevant literature for the images, a more accurate dataset can be constructed. Some or all of the above processing in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input the relevant literature data into a generating AI and have the generating AI perform the dataset accuracy improvement.

[0051] The system according to the embodiment is not limited to the example described above, and various modifications are possible, for example, as follows.

[0052] The CAPTCHA system can also include a feedback receiver to collect user feedback. This feedback receiver provides an interface for users to provide feedback on their annotations after they have completed them. For example, users can rate the difficulty and satisfaction level of their annotations. The feedback receiver can also improve the annotation interface and process based on user feedback. This allows for increased accuracy and efficiency of annotations by collecting user feedback and using it to improve the system.

[0053] The CAPTCHA system can also include a history analysis unit that analyzes the user's annotation history. For example, the history analysis unit analyzes the user's past annotation history to identify their strengths and weaknesses. If a user frequently performs annotations in a particular category, the history analysis unit will prioritize providing annotations related to that category. Furthermore, the history analysis unit can evaluate the accuracy of annotations based on the user's annotation history. This allows for improved annotation accuracy and efficiency by analyzing the user's annotation history.

[0054] The CAPTCHA system can further customize the annotation content based on the user's geographical location. For example, if the user is in a specific region, it can prioritize providing images related to that region. If the user is traveling, it can provide images related to their travel destination to capture their interest. Furthermore, if the user is at home, it can provide images related to their daily life. By customizing the annotation content based on the user's geographical location, it can attract user interest and improve the efficiency of annotation.

[0055] The CAPTCHA system can further customize the annotation content based on the user's social media activity. For example, if a user frequently posts about a particular topic on social media, it can provide images related to that topic. It can also provide images related to a specific hashtag if the user uses one. Furthermore, it can provide relevant images based on the posts of accounts the user follows. This allows for customized annotation content based on the user's social media activity, making it more engaging and improving the efficiency of the annotation process.

[0056] CAPTCHA systems can further customize annotation content based on the user's expertise and interests. For example, if a user is knowledgeable in a particular field, images related to that field will be prioritized. Similarly, if a user has a specific hobby, images related to that hobby can be provided. Furthermore, if a user works in a specific profession, images related to that profession can be provided. This customization of annotation content based on the user's expertise and interests improves the accuracy and efficiency of the annotation process.

[0057] The CAPTCHA system can further provide the optimal interface based on the user's past annotation history. For example, it can provide the optimal interface based on the interfaces the user has used in the past. It can also suggest the most efficient interface based on the user's past annotation history. Furthermore, it can analyze the user's past annotation history and provide the optimal interface. This improves the efficiency of annotation by providing the optimal interface based on the user's past annotation history.

[0058] The following briefly describes the processing flow for example form 1.

[0059] Step 1: The base labeling unit performs base labeling of the image. For example, it uses an AI agent to analyze the content of the image and assign an appropriate label. The AI ​​agent recognizes objects in the image and assigns a corresponding label to them. It can also perform labeling while considering the background information of the image. For example, if the background of the image contains specific cultural elements, those elements will be taken into consideration when labeling. Step 2: The annotation reception unit annotates the ambiguous parts of the images labeled by the base labeling unit. For example, a Japanese user performs the annotation via CAPTCHA. The user looks at the image displayed via CAPTCHA, recognizes its content, and applies the appropriate label. For example, they are required to apply the appropriate label to images of traditional Japanese festivals, food, landscapes, etc. Step 3: The dataset construction unit constructs a dataset specifically for Japanese culture based on the data annotated by the annotation reception unit. For example, it evaluates the annotation results and constructs a highly accurate dataset. The dataset construction unit clarifies the evaluation criteria and methods for the annotation results and improves the accuracy of the dataset.

[0060] (Example of form 2) The CAPTCHA system according to an embodiment of the present invention is a system in which Japanese users annotate Japanese culture. This CAPTCHA system solves the problem that conventional CAPTCHA systems mainly use images from overseas, which often felt unnatural to Japanese users. The CAPTCHA system aims to build a high-precision dataset specializing in Japanese culture and cultivate an AI model suitable for the Japanese market by having an AI agent perform base labeling and having Japanese users fill in ambiguous parts through CAPTCHA. For example, in the CAPTCHA system, the AI ​​agent first performs base labeling of images. Next, for ambiguous parts, Japanese users annotate through CAPTCHA. In this process, the user looks at images related to Japanese culture, recognizes their content, and labels them. For example, they are required to assign appropriate labels to images of traditional Japanese festivals, food, landscapes, etc. Through this mechanism, a high-precision dataset specializing in Japanese culture is efficiently collected. Furthermore, the AI ​​agent continuously learns and gains a deeper understanding of the unique characteristics of Japanese culture, enabling highly accurate annotation. As a result, an AI model suitable for the Japanese market is cultivated, and it is expected to have applications in various fields. For example, in the tourism industry, the development of AI models that provide information about Japanese tourist destinations and culture will enable the provision of more appropriate information to tourists. Furthermore, in the education sector, the development of AI models that automatically generate educational materials on Japanese history and culture is expected to improve the quality of education. Thus, by building high-precision datasets specifically for Japanese culture and cultivating AI models suited to the Japanese market, CAPTCHA systems are expected to have applications in various fields. This allows CAPTCHA systems to build high-precision datasets specifically for Japanese culture and cultivate AI models suited to the Japanese market.

[0061] The CAPTCHA system according to this embodiment comprises a base labeling unit, an annotation reception unit, and a dataset construction unit. The base labeling unit performs base labeling of images. The base labeling unit performs base labeling of images using, for example, an AI agent. The AI ​​agent can analyze the content of the image and assign appropriate labels. For example, the AI ​​agent can recognize objects in the image and assign corresponding labels to them. The AI ​​agent can also perform labeling while considering the background information of the image. For example, if the background of the image contains specific cultural elements, the labeling will take those elements into consideration. The annotation reception unit annotates ambiguous parts of the image that has been labeled by the base labeling unit. For example, the annotation reception unit receives annotations from Japanese users via CAPTCHA. The user looks at the image displayed via CAPTCHA, recognizes its content, and assigns appropriate labels. For example, they are required to assign appropriate labels to images of traditional Japanese festivals, food, landscapes, etc. The dataset construction unit constructs a dataset specializing in Japanese culture based on the data annotated by the annotation reception unit. The dataset construction unit, for example, evaluates the annotation results and constructs a high-precision dataset. The dataset construction unit clarifies the evaluation criteria and methods for the annotation results and improves the accuracy of the dataset. As a result, the CAPTCHA system according to this embodiment can construct a high-precision dataset specialized for Japanese culture and cultivate an AI model suitable for the Japanese market.

[0062] The base labeling unit performs base labeling of images. For example, the base labeling unit uses an AI agent to perform base labeling of images. The AI ​​agent can analyze the content of an image and assign appropriate labels. Specifically, the AI ​​agent uses deep learning technology to recognize objects within an image and assigns corresponding labels. For example, it recognizes objects unique to Japan, such as tatami rooms, shrine gates, and people wearing traditional Japanese clothing, and assigns labels such as "tatami," "torii," and "kimono" to them, respectively. The AI ​​agent can also consider the background information of the image when labeling. For example, if cherry blossoms are blooming in the background of the image, it will consider this element and assign labels such as "spring" and "cherry blossoms." Furthermore, in image analysis, the AI ​​agent utilizes object detection algorithms and semantic segmentation techniques to analyze each element within the image in detail. This allows the base labeling unit to analyze the content of images with high accuracy and assign appropriate labels. The AI ​​agent has been pre-trained on a large dataset of Japanese culture data, enabling it to recognize objects and background elements unique to Japan with high accuracy. This enables the base labeling unit to achieve highly accurate labeling specifically for Japanese culture, contributing to improved accuracy in subsequent annotation reception and dataset construction units.

[0063] The annotation reception unit annotates ambiguous parts of images labeled by the base labeling unit. For example, Japanese users perform annotation through CAPTCHA. Users view the image displayed through CAPTCHA, recognize its content, and assign appropriate labels. Specifically, users verify whether the labels assigned by the base labeling unit are correct for the displayed image and add corrections or additional labels as needed. For example, they are required to assign appropriate labels to images of traditional Japanese festivals, food, and landscapes. Users can consider details and background information within the image to assign more specific labels. For example, they might assign detailed labels such as "summer festival," "fireworks," and "yukata" to an image of a festival. The annotation reception unit also improves usability by designing the interface for users to perform annotation. For example, it provides features such as the ability to enlarge images and suggest label candidates to enable users to perform annotation efficiently. Furthermore, the annotation reception unit collects user annotation results in real time and stores them in a database. This allows the annotation reception unit to leverage the user's knowledge and experience to supplement ambiguous areas that the base labeling unit may have missed, thereby achieving highly accurate labeling.

[0064] The Dataset Construction Department builds datasets specifically for Japanese culture based on data annotated by the Annotation Reception Department. For example, the Dataset Construction Department evaluates annotation results to build highly accurate datasets. Specifically, it clarifies evaluation criteria and methods for annotation results to improve dataset accuracy. For instance, to evaluate the consistency and accuracy of annotation results, it compares annotation results from multiple users and selects reliable labels. Furthermore, the Dataset Construction Department introduces algorithms to automatically evaluate annotation results, efficiently building highly accurate datasets. In addition, the Dataset Construction Department regularly updates the constructed datasets to reflect the latest information. For example, it adds newly collected images and annotation results to enrich the dataset content. This ensures that the Dataset Construction Department can consistently provide highly accurate datasets based on the latest information. Finally, the Dataset Construction Department utilizes the constructed datasets to train other AI models, developing AI models suitable for the Japanese market. For example, it uses highly accurate datasets related to traditional Japanese festivals, food, and landscapes to train image recognition models and natural language processing models. This allows the dataset building unit to construct highly accurate datasets specifically for Japanese culture and cultivate AI models suitable for the Japanese market.

[0065] The base labeling unit can perform base labeling of images using an AI agent. For example, the base labeling unit can have the AI ​​agent analyze the content of the image and assign an appropriate label. For example, the AI ​​agent can recognize objects in the image and assign a corresponding label to them. The base labeling unit can also have the AI ​​agent consider the background information of the image when performing labeling. For example, if the background of the image contains specific cultural elements, the unit will take those elements into consideration when labeling. This improves the accuracy of base labeling by using an AI agent. Some or all of the above processing in the base labeling unit may be performed using a generative AI, or it may be performed without a generative AI. For example, the base labeling unit can input image data into a generative AI, analyze the content of the image, and perform labeling.

[0066] The annotation reception unit allows Japanese users to perform annotations via CAPTCHA. For example, the annotation reception unit can view images displayed by Japanese users through CAPTCHA, recognize their content, and assign appropriate labels. For instance, it is required to assign appropriate labels to images of traditional Japanese festivals, food, landscapes, etc. The annotation reception unit can also provide an interface for users to perform annotations. For example, it can provide a user-friendly screen layout and input method. This allows for the construction of a dataset specifically focused on Japanese culture through annotations by Japanese users. Some or all of the above-described processes in the annotation reception unit may be performed using AI, or not. For example, the annotation reception unit can input user input data into a generating AI, analyze the annotation content, and perform labeling.

[0067] The dataset construction unit can construct high-precision datasets specifically for Japanese culture. For example, the dataset construction unit constructs datasets based on data annotated by the annotation reception unit. The dataset construction unit evaluates the annotation results and constructs high-precision datasets. For example, the dataset construction unit clarifies the evaluation criteria and methods for annotation results to improve the accuracy of the dataset. By doing so, it is possible to cultivate AI models suitable for the Japanese market by constructing high-precision datasets specifically for Japanese culture. Some or all of the above-described processes in the dataset construction unit may be performed using AI, for example, or without using AI. For example, the dataset construction unit can input annotation data into a generation AI and have the generation AI execute the dataset construction.

[0068] The base labeling unit can continuously learn and gain a deeper understanding of the unique characteristics of Japanese culture. For example, the base labeling unit can set the update frequency of the training data and periodically learn new data. For example, the base labeling unit can collect new image data related to Japanese culture and add it to the training data. The base labeling unit can also optimize the learning algorithm to improve the efficiency of learning. For example, the base labeling unit can use a deep learning algorithm to more accurately capture image features. As a result, the accuracy of base labeling improves through continuous learning. Some or all of the above processes in the base labeling unit may be performed using AI, for example, or not using AI. For example, the base labeling unit can input training data into a generating AI and have the generating AI perform the optimization of the learning algorithm.

[0069] The annotation reception unit can provide an interface for users to perform annotations. For example, the annotation reception unit can provide a user-friendly screen layout and input method. For instance, it can provide a zoom function to make images easier for users to view and a dropdown menu to make label selection easier. The annotation reception unit can also provide guidelines and help functions for users to perform annotations. For example, it can provide guidelines that users can refer to when they get lost and a help function that explains how to operate the system. This improves the efficiency of annotation by providing an interface that makes it easy for users to perform annotations. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or not using AI. For example, the annotation reception unit can input user operation data into a generating AI and have the generating AI perform interface optimization.

[0070] The dataset construction unit can evaluate annotation results. For example, the dataset construction unit can clarify evaluation criteria and methods for annotation results to improve the accuracy of the dataset. For example, the dataset construction unit can evaluate the reliability of annotation results and prioritize the inclusion of highly reliable data in the dataset. The dataset construction unit can also evaluate the consistency of annotation results and include consistent data in the dataset. For example, the dataset construction unit can compare annotation results from multiple users and select consistent labels. This improves the accuracy of the dataset by evaluating the annotation results. Some or all of the above processes in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input annotation data into a generating AI and have the generating AI set evaluation criteria and perform the evaluation.

[0071] The base labeling unit can estimate the user's emotions and adjust the accuracy of the base labeling based on the estimated emotions. For example, if the user is stressed, the base labeling unit's AI agent can increase the accuracy of the labeling to reduce the user's burden. For example, if the user is relaxed, the base labeling unit's AI agent can slightly loosen the accuracy of the labeling and ask the user to do more annotation. Also, if the user is in a hurry, the base labeling unit's AI agent can maximize the accuracy of the labeling and minimize the user's annotation. In this way, the user's burden can be reduced by adjusting the accuracy of the base labeling based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or not using AI. For example, the base labeling unit can input user emotion data into the generating AI, allowing the AI ​​to perform accuracy adjustments for emotion-based labeling.

[0072] The base labeling unit can improve the accuracy of labeling based on the background information of the image. For example, if the background of the image contains specific cultural elements, the base labeling unit will take those elements into consideration when labeling. For example, if the background of the image is complex, the base labeling unit will have an AI agent analyze the background information to improve the accuracy of labeling. Also, if the background of the image is simple, the base labeling unit can have an AI agent focus on the main object and perform the labeling. This improves the accuracy of labeling by considering the background information of the image. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or without AI. For example, the base labeling unit can input image data into a generating AI, analyze the background information, and perform labeling.

[0073] The base labeling unit can adjust the level of detail in the labeling based on the date and time and location where the image was taken. For example, if the image was taken in a particular season, the base labeling unit will add a label related to that season. For example, if the image was taken in a particular location, the base labeling unit will add a label related to that location. Furthermore, if the date and time the image was taken is important, the base labeling unit can also adjust the level of detail in the labeling based on that date and time. This allows for more accurate labeling by adjusting the level of detail in the labeling based on the date and time and location where the image was taken. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or not using AI. For example, the base labeling unit can input the image metadata into a generating AI and have the generating AI perform the adjustment of the level of detail in the labeling based on the date and time and location where the image was taken.

[0074] The base labeling unit can estimate the user's emotions and determine the priority of labeling based on the estimated emotions. For example, if the user is stressed, the base labeling unit will prioritize high-importance labels. For example, if the user is relaxed, the base labeling unit will include low-importance labels as well. Furthermore, if the user is in a hurry, the base labeling unit can prioritize the most important labels. This reduces the user's burden by determining the priority of labeling based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or not using AI. For example, the base labeling unit can input user emotion data into the generative AI and have the generative AI perform emotion-based labeling priority determination.

[0075] The base labeling unit can perform labeling based on the color information of the image. For example, if the colors of the image have a specific cultural meaning, the base labeling unit will take that meaning into consideration when labeling. For example, if the colors of the image are vivid, the base labeling unit will perform labeling based on the dominant color. Also, if the colors of the image are muted, the base labeling unit can perform labeling by considering the overall color tone. This improves the accuracy of labeling by considering the color information of the image. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or without AI. For example, the base labeling unit can input image data into a generating AI, analyze the color information, and perform labeling.

[0076] The base labeling unit can improve the accuracy of labeling by analyzing the composition and elements of an image. For example, if the composition of an image has a specific cultural meaning, the base labeling unit will consider that meaning when labeling. For example, if the elements of an image are complex, the base labeling unit will focus on the main elements when labeling. Also, if the elements of an image are simple, the base labeling unit can consider the overall composition when labeling. In this way, the accuracy of labeling is improved by analyzing the composition and elements of an image. Some or all of the above processing in the base labeling unit may be performed using AI, for example, or without AI. For example, the base labeling unit can input image data into a generating AI, analyze the composition and elements, and perform labeling.

[0077] The annotation reception unit can estimate the user's emotions and adjust the annotation interface based on the estimated emotions. For example, if the user is tense, the annotation reception unit provides a simple and highly visible interface. For example, if the user is relaxed, the annotation reception unit provides an interface with detailed information. The annotation reception unit can also provide an interface that allows for quick annotation if the user is in a hurry. This improves the efficiency of annotation by adjusting the interface based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the annotation reception unit may be performed using AI or not using AI. For example, the annotation reception unit can input user emotion data into the generative AI and have the generative AI perform emotion-based interface adjustments.

[0078] The annotation reception unit can provide the optimal interface based on the user's past annotation history. For example, the annotation reception unit can provide the optimal interface based on the interface the user has used in the past. For example, the annotation reception unit can propose the most efficient interface from the user's past annotation history. The annotation reception unit can also analyze the user's past annotation history and provide the optimal interface. This improves the efficiency of annotation by providing the optimal interface based on the user's past annotation history. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without AI. For example, the annotation reception unit can input the user's annotation history data into a generating AI and have the generating AI perform the task of providing the optimal interface.

[0079] The annotation reception unit can customize the content of annotations based on the user's expertise and interests. For example, if a user is knowledgeable in a particular field, the annotation reception unit will prioritize providing annotations related to that field. For example, the annotation reception unit will provide relevant annotations based on the user's interests. The annotation reception unit can also provide optimal annotation content by considering the user's expertise. This improves the accuracy of annotations by customizing the content of annotations based on the user's expertise and interests. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without AI. For example, the annotation reception unit can input user profile data into a generating AI and have the generating AI perform the customization of the annotation content.

[0080] The annotation reception unit can estimate the user's emotions and determine annotation priorities based on those emotions. For example, if the user is stressed, the annotation reception unit will prioritize high-importance annotations. For example, if the user is relaxed, the annotation reception unit will also include less important annotations. Furthermore, if the user is in a hurry, the annotation reception unit can prioritize the most important annotations. This reduces the user's burden by determining annotation priorities based on their emotions. Emotion estimation is achieved using an emotion estimation function, such as an emotion engine or generative AI. Generative AI may be, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the annotation reception unit may be performed using AI or not. For example, the annotation reception unit can input user emotion data into a generative AI and have the generative AI perform emotion-based annotation prioritization.

[0081] The annotation reception unit can prioritize receiving annotations that are highly relevant based on the user's geographical location information. For example, if the user is in a specific region, the annotation reception unit will prioritize providing annotations related to that region. For example, the annotation reception unit will provide the most suitable annotations based on the user's geographical location information. The annotation reception unit can also prioritize receiving annotations that are highly relevant, taking into account the user's geographical location information. This improves the accuracy of annotations by prioritizing the acceptance of annotations that are highly relevant based on the user's geographical location information. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without using AI. For example, the annotation reception unit can input the user's geographical location data into a generating AI and have the generating AI perform the task of providing highly relevant annotations.

[0082] The annotation reception unit can accept relevant annotations based on the user's social media activity. For example, the annotation reception unit can provide annotations related to the user's areas of interest based on their social media activity. For example, the annotation reception unit can provide the most suitable annotations based on the user's social media activity. The annotation reception unit can also analyze the user's social media activity and prioritize accepting relevant annotations. This improves the accuracy of annotations by accepting relevant annotations based on the user's social media activity. Some or all of the above processing in the annotation reception unit may be performed using AI, for example, or without AI. For example, the annotation reception unit can input the user's social media data into a generating AI and have the generating AI perform the task of providing relevant annotations.

[0083] The dataset construction unit can estimate the user's emotions and adjust the dataset construction method based on the estimated user emotions. For example, if the user is relaxed, the dataset construction unit can construct a detailed dataset. For example, if the user is in a hurry, the dataset construction unit can construct a simplified dataset. Furthermore, if the user is excited, the dataset construction unit can construct a visually stimulating dataset. This reduces the user's burden by adjusting the dataset construction method based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the dataset construction unit may be performed using AI or not using AI. For example, the dataset construction unit can input user emotion data into the generative AI and have the generative AI perform the emotion-based adjustment of the dataset construction method.

[0084] The dataset construction unit can apply the optimal construction algorithm based on past datasets. For example, the dataset construction unit can select the optimal construction algorithm based on past datasets. For example, the dataset construction unit can apply an efficient construction algorithm from past datasets. The dataset construction unit can also apply the optimal construction algorithm by referring to past datasets. This improves the accuracy of the dataset by applying the optimal construction algorithm based on past datasets. Some or all of the above processes in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input past datasets into a generation AI and have the generation AI select and apply the optimal construction algorithm.

[0085] The dataset construction unit can apply different construction methods to each image category. For example, the dataset construction unit can select the optimal construction method for each image category. For example, the dataset construction unit can apply different construction methods depending on the image category. Alternatively, the dataset construction unit can apply the optimal construction method to each image category. By applying different construction methods to each image category, the accuracy of the dataset is improved. Some or all of the above processing in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input image data into a generation AI and have the generation AI execute the application of different construction methods for each category.

[0086] The dataset construction unit can estimate the user's emotions and determine the priority of datasets based on the estimated user emotions. For example, if the user is stressed, the dataset construction unit will prioritize building high-importance datasets. For example, if the user is relaxed, the dataset construction unit will also build datasets that are of lower importance. Furthermore, if the user is in a hurry, the dataset construction unit can also prioritize building the most important datasets. This reduces the user's burden by determining the priority of datasets based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the dataset construction unit may be performed using AI, for example, or not using AI. For example, the dataset construction unit can input user emotion data into a generative AI and have the generative AI perform emotion-based dataset prioritization.

[0087] The dataset construction unit can adjust the dataset structure based on the date, time, and location where the images were taken. For example, if an image was taken in a particular season, the dataset construction unit can construct a dataset related to that season. For example, if an image was taken in a particular location, the dataset construction unit can construct a dataset related to that location. Furthermore, if the date and time the images were taken is important, the dataset construction unit can also adjust the dataset structure based on that date and time. This allows for the construction of a more accurate dataset by adjusting the dataset structure based on the date, time, and location where the images were taken. Some or all of the above processing in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input image metadata into a generating AI and have the generating AI perform adjustments to the dataset structure based on the date, time, and location where the images were taken.

[0088] The dataset construction unit can improve the accuracy of the dataset based on the relevant literature for the images. For example, the dataset construction unit improves the accuracy of the dataset based on the relevant literature for the images. For example, the dataset construction unit constructs an optimal dataset by referring to the relevant literature for the images. The dataset construction unit can also improve the accuracy of the dataset based on the relevant literature for the images. By improving the accuracy of the dataset based on the relevant literature for the images, a more accurate dataset can be constructed. Some or all of the above processing in the dataset construction unit may be performed using AI, for example, or without AI. For example, the dataset construction unit can input the relevant literature data into a generating AI and have the generating AI perform the dataset accuracy improvement.

[0089] The system according to the embodiment is not limited to the example described above, and various modifications are possible, for example, as follows.

[0090] The CAPTCHA system can also include a feedback receiver to collect user feedback. This feedback receiver provides an interface for users to provide feedback on their annotations after they have completed them. For example, users can rate the difficulty and satisfaction level of their annotations. The feedback receiver can also improve the annotation interface and process based on user feedback. This allows for increased accuracy and efficiency of annotations by collecting user feedback and using it to improve the system.

[0091] The CAPTCHA system can also include a history analysis unit that analyzes the user's annotation history. For example, the history analysis unit analyzes the user's past annotation history to identify their strengths and weaknesses. If a user frequently performs annotations in a particular category, the history analysis unit will prioritize providing annotations related to that category. Furthermore, the history analysis unit can evaluate the accuracy of annotations based on the user's annotation history. This allows for improved annotation accuracy and efficiency by analyzing the user's annotation history.

[0092] The CAPTCHA system can also estimate the user's emotions and adjust the difficulty of annotations based on those emotions. For example, if a user is stressed, the system can provide easy annotations to reduce the user's burden. Conversely, if a user is relaxed, the system can provide more difficult annotations to improve the user's skills. Furthermore, if a user is in a hurry, the system can provide annotations that can be completed quickly. In this way, by adjusting the difficulty of annotations based on the user's emotions, the system reduces the user's burden and improves the efficiency of annotation.

[0093] The CAPTCHA system can further customize the annotation content based on the user's geographical location. For example, if the user is in a specific region, it can prioritize providing images related to that region. If the user is traveling, it can provide images related to their travel destination to capture their interest. Furthermore, if the user is at home, it can provide images related to their daily life. By customizing the annotation content based on the user's geographical location, it can attract user interest and improve the efficiency of annotation.

[0094] The CAPTCHA system can further customize the annotation content based on the user's social media activity. For example, if a user frequently posts about a particular topic on social media, it can provide images related to that topic. It can also provide images related to a specific hashtag if the user uses one. Furthermore, it can provide relevant images based on the posts of accounts the user follows. This allows for customized annotation content based on the user's social media activity, making it more engaging and improving the efficiency of the annotation process.

[0095] The CAPTCHA system can further estimate the user's emotions and customize the annotation interface based on those emotions. For example, if the user is nervous, it can provide a simple and highly visible interface. If the user is relaxed, it can provide an interface with more detailed information. Furthermore, if the user is in a hurry, it can provide an interface that allows for quick annotation. By customizing the annotation interface based on the user's emotions, the system reduces the user's burden and improves the efficiency of annotation.

[0096] CAPTCHA systems can further customize annotation content based on the user's expertise and interests. For example, if a user is knowledgeable in a particular field, images related to that field will be prioritized. Similarly, if a user has a specific hobby, images related to that hobby can be provided. Furthermore, if a user works in a specific profession, images related to that profession can be provided. This customization of annotation content based on the user's expertise and interests improves the accuracy and efficiency of the annotation process.

[0097] The CAPTCHA system can further estimate the user's emotions and prioritize annotations based on those emotions. For example, if the user is stressed, it will prioritize high-priority annotations. If the user is relaxed, it can also include lower-priority annotations. Furthermore, if the user is in a hurry, it can prioritize the most important annotations. By prioritizing annotations based on the user's emotions, this reduces the user's burden and improves annotation efficiency.

[0098] The CAPTCHA system can further provide the optimal interface based on the user's past annotation history. For example, it can provide the optimal interface based on the interfaces the user has used in the past. It can also suggest the most efficient interface based on the user's past annotation history. Furthermore, it can analyze the user's past annotation history and provide the optimal interface. This improves the efficiency of annotation by providing the optimal interface based on the user's past annotation history.

[0099] The CAPTCHA system can further estimate the user's emotions and adjust how the dataset is constructed based on those emotions. For example, if the user is relaxed, a detailed dataset can be constructed. If the user is in a hurry, a simplified dataset can be constructed. Furthermore, if the user is excited, a visually stimulating dataset can be constructed. By adjusting how the dataset is constructed based on the user's emotions, the system reduces the user's burden and improves the accuracy of the dataset.

[0100] The following briefly describes the processing flow for example form 2.

[0101] Step 1: The base labeling unit performs base labeling of the image. For example, it uses an AI agent to analyze the content of the image and assign an appropriate label. The AI ​​agent recognizes objects in the image and assigns a corresponding label to them. It can also perform labeling while considering the background information of the image. For example, if the background of the image contains specific cultural elements, those elements will be taken into consideration when labeling. Step 2: The annotation reception unit annotates the ambiguous parts of the images labeled by the base labeling unit. For example, a Japanese user performs the annotation via CAPTCHA. The user looks at the image displayed via CAPTCHA, recognizes its content, and applies the appropriate label. For example, they are required to apply the appropriate label to images of traditional Japanese festivals, food, landscapes, etc. Step 3: The dataset construction unit constructs a dataset specifically for Japanese culture based on the data annotated by the annotation reception unit. For example, it evaluates the annotation results and constructs a highly accurate dataset. The dataset construction unit clarifies the evaluation criteria and methods for the annotation results and improves the accuracy of the dataset.

[0102] 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.

[0103] Data generation model 58 is a form of 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> Examples of generative AI include text generation AI, image generation AI, and multimodal generation AI. 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 (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats from audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVMs), k-means clustering, convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each of the above parts is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example.Furthermore, processing performed by AI, including generative AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by AI, including generative AI.

[0104] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart device 14 or an external device, and the smart device 14 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0105] Each of the multiple elements described above, including the base labeling unit, annotation reception unit, and dataset construction unit, is implemented in at least one of the smart device 14 and the data processing device 12. For example, the base labeling unit is implemented by the processor 46 of the smart device 14 and performs base labeling of images using an AI agent. The annotation reception unit is implemented by the control unit 46A of the smart device 14 and allows the user to perform annotations via CAPTCHA. The dataset construction unit is implemented by the identification processing unit 290 of the data processing device 12 and evaluates the annotation results to construct a high-precision dataset. The correspondence between each unit and the device or control unit is not limited to the example described above and can be modified in various ways.

[0106] [Second Embodiment] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0107] 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.

[0108] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0109] 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.

[0110] The microphone 238 receives voice signals from the user and accepts instructions from the user. The microphone 238 captures the voice signals from the user, 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.

[0111] 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, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0112] 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.

[0113] 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 by the processor 28. The storage 32 stores the specific processing program 56.

[0114] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0115] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0116] In the smart glasses 214, specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 acting as a control unit 46A according to the specific processing program 60 executed on the RAM 48. The smart glasses 214 also have a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0117] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0118] 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.

[0119] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. 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 inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0120] The data processing system 210 according to the second embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 210 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart glasses 214 or an external device, and the smart glasses 214 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0121] Each of the multiple elements described above, including the base labeling unit, annotation reception unit, and dataset construction unit, is implemented in at least one of the smart glasses 214 and the data processing device 12. For example, the base labeling unit is implemented by the processor 46 of the smart glasses 214 and performs base labeling of images using an AI agent. The annotation reception unit is implemented by the control unit 46A of the smart glasses 214 and allows the user to perform annotations via CAPTCHA. The dataset construction unit is implemented by the identification processing unit 290 of the data processing device 12 and evaluates the annotation results to construct a high-precision dataset. The correspondence between each unit and the device or control unit is not limited to the example described above and can be modified in various ways.

[0122] [Third Embodiment] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0123] 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.

[0124] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0125] 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.

[0126] The microphone 238 receives voice signals from the user and accepts instructions from the user. The microphone 238 captures the voice signals from the user, 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.

[0127] 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, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0128] 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.

[0129] 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.

[0130] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0131] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0132] In the headset terminal 314, specific processing is performed by the processor 46. The storage 50 stores a specific program 60. The processor 46 reads the specific program 60 from the storage 50 and executes the read specific program 60 on the RAM 48. The specific processing is realized by the processor 46 acting as a control unit 46A according to the specific program 60 executed on the RAM 48. The headset terminal 314 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0133] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0134] 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.

[0135] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. 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 inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0136] The data processing system 310 according to the third embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 310 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the headset terminal 314, but may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the headset terminal 314. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the headset terminal 314 or an external device, and the headset terminal 314 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0137] Each of the multiple elements described above, including the base labeling unit, annotation reception unit, and dataset construction unit, is implemented in at least one of the headset terminal 314 and the data processing device 12. For example, the base labeling unit is implemented by the processor 46 of the headset terminal 314 and performs base labeling of images using an AI agent. The annotation reception unit is implemented by the control unit 46A of the headset terminal 314 and allows the user to perform annotations via CAPTCHA. The dataset construction unit is implemented by the identification processing unit 290 of the data processing device 12 and evaluates the annotation results to construct a high-precision dataset. The correspondence between each unit and the device or control unit is not limited to the example described above and can be modified in various ways.

[0138] [Fourth Embodiment] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0139] 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.

[0140] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0141] 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.

[0142] The microphone 238 receives voice signals from the user and accepts instructions from the user. The microphone 238 captures the voice signals from the user, 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.

[0143] 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 image sensor or CCD image sensor, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0144] 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.

[0145] 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. The robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0146] 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.

[0147] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0148] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0149] In robot 414, specific processing is performed by processor 46. A specific program 60 is stored in storage 50. Processor 46 reads the specific program 60 from storage 50 and executes it on RAM 48. The specific processing is achieved by processor 46 acting as a control unit 46A according to the specific program 60 executed on RAM 48. Robot 414 also has data generation model 58 and emotion identification model 59, similar to those of the robot, and can perform processing similar to that of the specific processing unit 290 using these models.

[0150] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0151] 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.

[0152] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. 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 inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0153] The data processing system 410 according to the fourth embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 410 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the robot 414 or an external device, and the robot 414 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0154] Each of the multiple elements described above, including the base labeling unit, annotation reception unit, and dataset construction unit, is implemented in at least one of the robot 414 and the data processing unit 12. For example, the base labeling unit is implemented by the processor 46 of the robot 414 and performs base labeling of images using an AI agent. The annotation reception unit is implemented by the control unit 46A of the robot 414 and allows the user to perform annotations via CAPTCHA. The dataset construction unit is implemented by the identification processing unit 290 of the data processing unit 12 and evaluates the annotation results to construct a high-precision dataset. The correspondence between each unit and the device or control unit is not limited to the example described above and can be modified in various ways.

[0155] 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.

[0156] Figure 9 shows the 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.

[0157] 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.

[0158] 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.

[0159] 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, and motorcycles, 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 based, for example, 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.

[0160] 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."

[0161] 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.

[0162] 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 method for the specific process may be used, which includes computer 22 and multiple other computers.

[0163] 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.

[0164] 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.

[0165] 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.

[0166] 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.

[0167] 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.

[0168] 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.

[0169] 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.

[0170] Furthermore, although the above-described examples were divided into four embodiments, some or all of these embodiments may be combined. Also, the smart device 14, smart glasses 214, headset terminal 314, and robot 414 are just examples, and they may be combined, or other devices may be used. Also, although the above-described examples were divided into two embodiments, Embodiment 1 and Embodiment 2, these may be combined.

[0171] 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 other things 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.

[0172] 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.

[0173] (Note 1) A base labeling unit that performs base labeling on images, An annotation reception unit that annotates ambiguous parts of the image labeled by the base labeling unit, The system comprises a dataset construction unit that constructs a dataset specifically for Japanese culture based on the data annotated by the annotation reception unit. A system characterized by the following features. (Note 2) The base labeling section is, Perform image base labeling using an AI agent. The system described in Appendix 1, characterized by the features described herein. (Note 3) The annotation reception unit is, Japanese users perform annotations via CAPTCHA. The system described in Appendix 1, characterized by the features described herein. (Note 4) The aforementioned dataset construction unit, Building a high-precision dataset specifically focused on Japanese culture. The system described in Appendix 1, characterized by the features described herein. (Note 5) The base labeling section is, Continue learning to gain a deeper understanding of the unique characteristics of Japanese culture. The system described in Appendix 1, characterized by the features described herein. (Note 6) The annotation reception unit is, Provides an interface for users to perform annotations. The system described in Appendix 1, characterized by the features described herein. (Note 7) The aforementioned dataset construction unit, Evaluate the annotation results. The system described in Appendix 1, characterized by the features described herein. (Note 8) The base labeling section is, It estimates the user's emotions and adjusts the accuracy of base labeling based on the estimated user emotions. The system described in Appendix 1, characterized by the features described herein. (Note 9) The base labeling section is, Improve labeling accuracy based on image background information. The system described in Appendix 1, characterized by the features described herein. (Note 10) The base labeling section is, Adjust the level of detail in the labeling based on the date and location where the image was taken. The system described in Appendix 1, characterized by the features described herein. (Note 11) The base labeling section is, It estimates the user's emotions and determines the labeling priority based on the estimated user emotions. The system described in Appendix 1, characterized by the features described herein. (Note 12) The base labeling section is, Labeling is performed based on the color information of the image. The system described in Appendix 1, characterized by the features described herein. (Note 13) The base labeling section is, Analyze the composition and elements of the image to improve the accuracy of labeling. The system described in Appendix 1, characterized by the features described herein. (Note 14) The annotation reception unit is, It estimates the user's emotions and adjusts the annotation interface based on those estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 15) The annotation reception unit is, Provides the optimal interface based on the user's past annotation history. The system described in Appendix 1, characterized by the features described herein. (Note 16) The annotation reception unit is, Customize annotation content based on the user's expertise and interests. The system described in Appendix 1, characterized by the features described herein. (Note 17) The annotation reception unit is, The system estimates the user's emotions and determines annotation priorities based on those estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 18) The annotation reception unit is, The system prioritizes accepting annotations that are highly relevant based on the user's geographical location. The system described in Appendix 1, characterized by the features described herein. (Note 19) The annotation reception unit is, Accepts relevant annotations based on users' social media activity. The system described in Appendix 1, characterized by the features described herein. (Note 20) The aforementioned dataset construction unit, We estimate user sentiment and adjust how the dataset is constructed based on the estimated user sentiment. The system described in Appendix 1, characterized by the features described herein. (Note 21) The aforementioned dataset construction unit, Apply the optimal construction algorithm based on past datasets. The system described in Appendix 1, characterized by the features described herein. (Note 22) The aforementioned dataset construction unit, Apply different construction methods to each image category. The system described in Appendix 1, characterized by the features described herein. (Note 23) The aforementioned dataset construction unit, It estimates user sentiment and prioritizes the dataset based on the estimated user sentiment. The system described in Appendix 1, characterized by the features described herein. (Note 24) The aforementioned dataset construction unit, Adjust the dataset structure based on the date, time, and location where the images were taken. The system described in Appendix 1, characterized by the features described herein. (Note 25) The aforementioned dataset construction unit, Improve the accuracy of the dataset based on related literature for the images. The system described in Appendix 1, characterized by the features described herein. [Explanation of Symbols]

[0174] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots

Claims

1. A base labeling unit that performs base labeling on images, An annotation reception unit that annotates ambiguous parts of the image labeled by the base labeling unit, The system comprises a dataset construction unit that constructs a dataset specifically for Japanese culture based on the data annotated by the annotation reception unit. A system characterized by the following features.

2. The base labeling section is, Perform image base labeling using an AI agent. The system according to feature 1.

3. The annotation reception unit is, Japanese users perform annotations via CAPTCHA. The system according to feature 1.

4. The aforementioned dataset construction unit, Building a high-precision dataset specifically focused on Japanese culture. The system according to feature 1.

5. The base labeling section is, Continue learning to gain a deeper understanding of the unique characteristics of Japanese culture. The system according to feature 1.

6. The annotation reception unit is, Provides an interface for users to perform annotations. The system according to feature 1.

7. The aforementioned dataset construction unit, Evaluate the annotation results. The system according to feature 1.

8. The base labeling section is, It estimates the user's emotions and adjusts the accuracy of base labeling based on the estimated user emotions. The system according to feature 1.

Images