system

Abstract

Provide a system. [Solution] a means for presenting a plurality of facial images to a user; a means for a user to select from the displayed facial images; means of collecting data based on user preferences; A means of learning user preferences based on collected data; and A means for generating a new face image that reflects the learned features; A system including:

Description

[Technical Field]

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

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

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional image generation AI requires prompt input when generating a facial image of a user's choice. However, it is difficult to verbalize facial features, which requires a great deal of effort from the user. Furthermore, simply changing the model or settings cannot reflect the user's specific preferences, making it difficult to obtain highly accurate results. For this reason, there is a demand for a system that can be operated intuitively by the user and that can efficiently generate facial images of the user's choice. [Means for solving the problem]

[0005] The present invention provides a means for a user to select a preferred facial image from multiple facial images and a means for learning the user's preferences based on the selection information. Specifically, the system includes a means for presenting multiple facial images to a user, a means for the user to select from the displayed facial images, a means for collecting data based on the user's selection, a means for learning the user's preferred features based on the collected data, and a means for generating a new facial image that reflects the learned features. This system allows the user to avoid prompt input and efficiently generate a preferred facial image intuitively. Furthermore, the system also includes a means for gradually generating a preferred facial image by repeating the user's selection, and a means for the user to fine-tune the generated facial image, thereby achieving a more satisfactory result for the user.

[0006] "User" refers to the person who operates the system to select and confirm preferred facial images.

[0007] "Facial Images" refers to the facial images generated by the system and presented to the user.

[0008] "Means" refers to elements such as methods, devices, and programs for realizing each function of this system.

[0009] "Means for presenting" refers to the method or function for transmitting the facial image generated from the server to the terminal and displaying it to the user.

[0010] "Means for selection" refers to a method or function that allows a user to select a preferred face from the presented face images.

[0011] "Means of collection" refers to the method or function for recording user selection information and transmitting it to the server.

[0012] "Means of learning" refers to the methods and functions by which AI analyzes user preferences and learns their characteristics based on collected user selection information.

[0013] "Means for generation" refers to the method or function for generating new facial images that reflect the learned user's preferred features.

[0014] "Iterative" refers to the user selection and learning / generation process being repeated multiple times.

[0015] "Fine-tuning" refers to the user making changes or checking details of the final generated facial image. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12]FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0017] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

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

[0019] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0020] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0021] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0022] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0023] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0024] [First embodiment]

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

[0026] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0027] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0028] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0029] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0030] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0031] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

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

[0033] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0034] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0035] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0036] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0037] The present invention provides a system for efficiently generating facial images that suit a user's preferences. The system includes a series of processes that present multiple facial images, learn the user's preferences based on the user's selection, and generate new facial images.

[0038] System configuration

[0039] server

[0040] It has an image generation AI model.

[0041] An initial face image and a new face image are generated.

[0042] It receives user selection information and learns based on that information.

[0043] Terminal

[0044] The device the user operates (e.g., PC, smartphone).

[0045] The facial image sent from the server is displayed to the user.

[0046] The information of the face image selected by the user is sent to the server.

[0047] user

[0048] Select your preferred facial image and provide feedback to the system.

[0049] Processing flow

[0050] Presentation of initial face image

[0051] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device.

[0052] The device displays these facial images to the user.

[0053] User Selection

[0054] The user selects the facial image that he or she likes best from the presented facial images.

[0055] The terminal transmits this selection information to the server.

[0056] Learning preferences

[0057] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[0058] Generating new face images

[0059] The server generates the next stage of facial images based on the learning results and sends them to the terminal again.

[0060] A new face image will be displayed that reflects the user's selection.

[0061] Iterative Process

[0062] The user again selects the face image they like best from the new set of face images, repeating this process multiple times to provide feedback to the system.

[0063] Through this process, the server gradually learns the user's preferences and generates highly accurate facial images.

[0064] Specific examples

[0065] 1. Presenting the initial face image:

[0066] A group of face images (e.g., A1, A2, A3, A4, A5) generated by the server are displayed on the terminal.

[0067] The user selects A3 and the device sends the information to the server.

[0068] 2. Learning preferences:

[0069] The server extracts features from A3's facial images and learns its preferences.

[0070] 3. Generate a new face image:

[0071] The server generates a new set of facial images (e.g., B1, B2, B3, B4, B5) based on the learning results and sends them to the terminal.

[0072] The user again selects B2, which they like best, and the process is repeated.

[0073] 4. Generate the final face image:

[0074] Finally, after multiple rounds of selection and learning, the server generates a facial image that best reflects the user's preferences and sends it to the device.

[0075] The user can review this facial image and make minor adjustments as needed.

[0076] This system allows users to avoid the hassle of prompt input and use intuitive operations to accurately generate their preferred facial images.

[0077] The processing flow will be explained below.

[0078] Step 1:

[0079] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[0080] Step 2:

[0081] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[0082] Step 3:

[0083] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[0084] Step 4:

[0085] The server receives the user's selection data, extracts features from the selected facial image (e.g., eye shape, nose height, skin color, etc.), and trains an AI model based on the feature information, thereby learning the user's preferences.

[0086] Step 5:

[0087] The server generates a new set of facial images that reflect the learned user's preferred features. These facial images are adjusted to more closely match the user's preferences. The newly generated set of facial images is then sent to the device.

[0088] Step 6:

[0089] The device will then display the newly sent face images to the user again, ready for the user to select the face image they like best.

[0090] Step 7:

[0091] The user again selects one of the newly presented face images that he or she likes best. This selection information is also recorded on the device and sent to the server. As in step 4, the user's selection information is again learned by the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0092] Step 8:

[0093] After multiple selections and learning processes are completed, the server finally generates a facial image that best reflects the user's preferences, and sends this final facial image to the device.

[0094] Step 9:

[0095] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[0096] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input and allowing the system to learn step-by-step based on user selection, highly accurate results can be achieved.

[0097] Example 1

[0098] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0099] Conventional facial image generation systems have struggled to efficiently generate the facial image desired by the user, often forcing the user to perform a huge amount of input work. Furthermore, due to a lack of systems that accurately respond to the user's preferences, the generated facial image often does not meet the user's expectations. The objective of this invention is to solve these problems and provide a system that enables users to intuitively operate the system and generate their preferred facial image with high accuracy.

[0100] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0101] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for transmitting the generated facial image to a user terminal, means for displaying the transmitted facial image to the user terminal, and means for the user to select a preferred facial image from the displayed new facial images. This allows the user to avoid the hassle of prompt input and to generate a preferred facial image with high accuracy through intuitive operation.

[0102] "User" refers to an individual who uses the System to select and generate their preferred facial image.

[0103] "Server" refers to a computer system equipped with an image generation AI model that generates and learns facial images based on user selection information.

[0104] "User terminal" refers to a device operated by a user (e.g., a PC or smartphone) that displays the facial image sent from the server and transmits the user's selection to the server.

[0105] "Image generation AI model" refers to an artificial intelligence model (e.g., StyleGAN) that generates facial images using deep learning technology.

[0106] "Facial Image" means an image of a human face that is generated by the system and presented to the user.

[0107] "Selection information" refers to data (such as the image ID and features) generated when a user selects one face image from multiple images.

[0108] "Features" refer to visual elements contained in a facial image (e.g., eye shape, nose height, skin color, etc.).

[0109] "Learning" refers to the process by which the AI ​​model analyzes and incorporates the user's preferences based on the information in the facial image selected by the user.

[0110] "Display" refers to the act of a user device providing a facial image to the user's visual sense.

[0111] "Generation" refers to the process by which an image-generating AI model generates new facial images.

[0112] "Fine-tuning" refers to the operation in which the user modifies the details of the final generated facial image to make it more closely match their own preferences.

[0113] This invention is a system for efficiently generating facial images that suit a user's preferences. This system consists of several main components, each of which performs a specific function. Specifically, it consists of a server, a user terminal, and a generative AI model.

[0114] System configuration

[0115] server

[0116] The server is equipped with an image generation AI model and has the following functions:

[0117] 1. Generate initial face images: Use an image generation AI model (e.g., StyleGAN) to generate a set of initial face images.

[0118] 2. Learning function: Based on the user's selection information, the system analyzes the characteristics of the facial image and learns the user's preferences.

[0119] 3. Generate new face images: The learned features are reflected to generate the next set of face images.

[0120] 4. Sending face image: Send the generated face image to the user terminal.

[0121] User Device

[0122] The user terminal is a device (e.g., PC, smartphone) that receives face images from the server and is operated by the user. It has the following functions:

[0123] 1. Displaying face image: The face image sent from the server is displayed to the user.

[0124] 2. Sending selected information: Send the information of the face image selected by the user to the server.

[0125] user

[0126] Users interact with the system as follows:

[0127] 1. Select a face image: Select the face image you like best from the displayed face images.

[0128] 2. Providing feedback: Giving feedback to the system through multiple selections.

[0129] Hardware and software used

[0130] Hardware: Servers equipped with high-performance GPUs, devices operated by users (PCs, smartphones)

[0131] Software: Generative AI model (e.g., StyleGAN), user interface (web application or mobile app)

[0132] Specific examples

[0133] 1. The server generates a set of initial face images using StyleGAN. For example, five face images (e.g., A1, A2, A3, A4, A5) are generated.

[0134] 2. The server sends the generated initial face images to the user's device.

[0135] 3. The user selects the facial image (e.g., A3) that they like best from the initial facial images displayed on the device. The device then sends this selection information to the server.

[0136] 4. The server analyzes the features of the selected face image (A3) and learns the user's preferences.

[0137] 5. The server generates a new set of face images (e.g., B1, B2, B3, B4, B5) reflecting the learning results and sends them again to the user's device.

[0138] 6. The user again selects the most preferred face image (e.g., B2) from the new set of face images, and the same process is repeated.

[0139] Prompt Sentence Examples

[0140] Design a system that allows a user to select their first preferred facial image and then learns its features to generate new facial images.

[0141] This system allows users to avoid the hassle of prompt input and to intuitively generate their preferred facial images with high accuracy.

[0142] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0143] Step 1:

[0144] The server generates a set of initial face images using an image generation AI model (e.g., StyleGAN), which uses a random latent space vector as input and generates multiple face images as output. Through this process, the system can obtain an initial set of face images (e.g., A1, A2, A3, A4, A5) to present to the user.

[0145] Step 2:

[0146] The server then sends the generated initial face images to the user's device. The generated face image data is used as input, and is sent to the user's device as output. This process enables the initial face images to be displayed on the device operated by the user.

[0147] Step 3:

[0148] The terminal displays the initial face images received from the server to the user. The received face image data is used as input and displayed on the user's screen as output. This process allows the user to check and select from the initial face images.

[0149] Step 4:

[0150] The user selects the face image they like best from the initial face images displayed. The visually displayed face image is used as input, and the ID and feature information of the selected face image are generated as output. This process determines the user's selection information.

[0151] Step 5:

[0152] The terminal sends the selection information of the face image selected by the user to the server. The user's selection information is used as input and sent to the server as output. This process allows the server to receive the user's preferences.

[0153] Step 6:

[0154] The server analyzes the features of the selected facial image based on the received selection information. The selected facial image data is used as input, and a feature vector is generated as output. This process provides the server with information to learn the user's preferred features.

[0155] Step 7:

[0156] The server inputs the analyzed feature vectors into a generative AI model to learn the user's preferences. The feature vectors are used as input, and updated model parameters are obtained as output. This process allows the generative AI model to reflect the user's preferences.

[0157] Step 8:

[0158] The server then generates a new set of facial images based on the learning results. The updated generative AI model is used as input, and new facial images (e.g., B1, B2, B3, B4, B5) are generated as output. This process provides the next phase of facial images to the user.

[0159] Step 9:

[0160] The server then sends the newly generated face images back to the user's device. The newly generated face image data is used as input, and is sent to the user's device as output. This process allows the user to check the new face images.

[0161] Step 10:

[0162] The device then presents a new set of face images to the user, using the new face image data sent from the server as input and displaying it on the user's screen as output. This process allows the user to select a face image again.

[0163] Step 11:

[0164] The user then selects the face image they like best from a set of new face images. The new face image displayed is used as input, and the information about the reselected face image is generated as output. This process is repeated multiple times until the face image closest to the user's preference is generated.

[0165] (Application example 1)

[0166] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0167] Conventional facial image generation systems have difficulty efficiently generating facial images that match a user's preferences. Furthermore, when applying the generated facial images to a virtual store clerk character, the user's preferences are not fully reflected. Therefore, improving the user experience has become a challenge.

[0168] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0169] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for presenting the generated facial images to the user in a virtual store and for the user to select one, and means for applying the newly generated facial image to a store clerk character in the virtual store. This allows the user to intuitively generate a preferred facial image and have it reflected in the store clerk character in the virtual store.

[0170] A "user" is a person who operates the system, selects his / her preferred facial image, and provides the system with the selected information.

[0171] A "face image" is an image with facial features that is presented for selection by the user.

[0172] The "presentation means" is a mechanism for displaying the facial image generated by the server on the terminal operated by the user.

[0173] The "means for making a selection" is a mechanism that provides an interface for a user to select a preferred facial image from a plurality of displayed facial images.

[0174] "Means for collecting data" refers to a mechanism for transmitting user selection information to a server and storing it.

[0175] The "means of learning features" is a mechanism that uses AI models to analyze and learn users' facial image preferences based on collected user selection data.

[0176] The "means for generating a new facial image" is a mechanism for generating a new facial image that reflects the user's preferences based on the learning results.

[0177] A "virtual store" is a store that exists on the Internet or in a virtual space and allows users to browse and purchase products online.

[0178] A "store clerk character" is a virtual character that appears in the virtual store to introduce and recommend products.

[0179] The present invention provides a system that allows users to efficiently generate facial images of their choice and apply the generated facial images to store clerk characters in a virtual store. This system operates mainly on a server and terminals and includes a process for learning and reflecting user preferences.

[0180] System configuration

[0181] server

[0182] It is equipped with an image generation AI model that generates initial and new facial images, and also receives user selection information and learns based on that information.

[0183] Terminal

[0184] A device operated by the user (e.g., PC, smartphone). It displays the facial images sent from the server to the user and sends the information of the facial image selected by the user to the server.

[0185] user

[0186] Select your preferred facial image and provide feedback to the system.

[0187] Processing flow

[0188] Presentation of initial face image

[0189] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device, where the user is shown these facial images.

[0190] User Selection

[0191] The user selects the facial image that they like best from the presented facial images, and the terminal transmits this selection information to the server.

[0192] Learning preferences

[0193] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[0194] Generating new face images

[0195] The server generates the next set of facial images based on the learning results and sends them back to the device, where new facial images reflecting the user's selection are displayed.

[0196] Applies to virtual stores

[0197] The generated facial image is applied to a store clerk character in a virtual store, and the user can interact with this store clerk character in the virtual store.

[0198] Hardware and software used

[0199] Hardware: Smartphone (iOS or ANDROID (registered trademark)), cloud server

[0200] Software: Python, PIL (Python Imaging Library), tkinter (for creating GUI)

[0201] Data processing and calculation

[0202] 1. Presentation of initial face images: Obtain the face image group URL generated by the server, download each image, and process it into display data.

[0203] 2. Sending user selection information: Display a dialog using tkinter, get the index of the image selected by the user, and send it to the server.

[0204] 3. Generate new face images: The server obtains the new face image URL and downloads each image.

[0205] 4. Iteration: The basic process is repeated until a facial image that satisfies the user is generated.

[0206] 5. Application in virtual stores: By applying the generated facial image to a clerk character in a virtual store, users can interact with their favorite clerk character.

[0207] Examples of concrete examples and prompts

[0208] Presentation of initial face image

[0209] The user selects the image they like best from the first five facial images displayed (e.g., A1, A2, A3, A4, A5).

[0210] Prompt: "Please choose the face image below that you like best."

[0211] Sending user selection information

[0212] For example, if the user selects A3, this information is sent to the server.

[0213] Prompt: "A3 selected. Generate a new face image."

[0214] Generating new face images

[0215] The server generates a set of new face images (e.g., B1, B2, B3, B4, B5) and presents them to the user again.

[0216] Prompt: "Facial images for the next step have been generated. Please choose the one you like best again."

[0217] In this way, the user can intuitively create a facial image of their choice and apply it to a salesperson character in a virtual store.

[0218] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0219] Step 1:

[0220] The server uses an image generation AI model to generate a set of initial facial images and sends their URLs to the device.

[0221] Input: A request comes to the server to generate the initial image.

[0222] Data processing / data calculation: The server generates initial facial images using an image generation AI model, generates a URL for each facial image, and sends it to the device.

[0223] Output: URL list of initial face images.

[0224] Step 2:

[0225] The device downloads the image using the URL of the initial face image received from the server and displays it to the user.

[0226] Input: URL list of initial face images sent from the server.

[0227] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[0228] Output: Multiple initial face images shown to the user.

[0229] Step 3:

[0230] The user selects the facial image that he or she likes best from the presented facial images.

[0231] Input: Multiple displayed face images.

[0232] Data processing / data calculation: Users intuitively select the facial image they like best.

[0233] Output: The index of the selected face image.

[0234] Step 4:

[0235] The terminal transmits the index of the face image selected by the user to the server.

[0236] Input: The index of the face image selected by the user.

[0237] Data processing / data calculation: Converts index information into data to be sent to the server.

[0238] Output: Sends the selection index to the server.

[0239] Step 5:

[0240] The server extracts features from the selected facial image and inputs them into an image generation AI model to learn the user's preferences.

[0241] Input: The index of the selected face image.

[0242] Data processing / data calculation: The server extracts the features of the selected facial image (eye shape, nose height, skin color, etc.) and uses an AI model to learn the user's preferences.

[0243] Output: The training results.

[0244] Step 6:

[0245] The server generates a new set of facial images based on the learning results and sends the URL to the terminal.

[0246] Input: Training results.

[0247] Data processing / data calculation: The server generates new facial images using an AI model and sends their URLs to the device.

[0248] Output: A list of URLs for new face images.

[0249] Step 7:

[0250] The device will use the new face image URL to download the image and display it to the user again.

[0251] Input: A list of URLs for new face images.

[0252] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[0253] Output: Multiple new face images that are displayed to the user.

[0254] Step 8:

[0255] Repeat steps 3 through 7 until the user is satisfied with the generation process.

[0256] Input: A new face image.

[0257] Data processing / data calculation: The user again selects the facial image they like best, and a new facial image is generated based on that.

[0258] Output: The final selected face image.

[0259] Step 9:

[0260] The generated final facial image is applied to a clerk character in a virtual store.

[0261] Input: The final selected face image.

[0262] Data processing / data calculation: The server converts the selected facial image into data to be applied to the virtual store clerk character.

[0263] Output: A custom store clerk character to be used in the virtual store.

[0264] Step 10:

[0265] Users interact with store clerk characters generated within the virtual store.

[0266] Input: A virtual store where the clerk character applies.

[0267] Data processing / data calculation: A store clerk character behaves within the virtual store and interacts with the user.

[0268] Output: A user experience interacting with a custom store clerk character.

[0269] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0270] This invention is a system for efficiently generating facial images that suit a user's preferences, and also combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes that present multiple facial images, learn the user's preferences based on the user's selection and the emotions at that time, and generate new facial images.

[0271] System configuration

[0272] server

[0273] It has an image generation AI model.

[0274] An initial face image and a new face image are generated.

[0275] It receives user selection information and emotion data and learns based on that information.

[0276] Terminal

[0277] The device the user operates (e.g., PC, smartphone).

[0278] The facial image sent from the server is displayed to the user.

[0279] Information about the face image selected by the user and emotion data obtained from the emotion engine are sent to the server.

[0280] user

[0281] Select your preferred facial image and provide feedback to the system.

[0282] Emotion data is collected by an emotion engine.

[0283] Emotion Engine

[0284] It recognizes emotions from the user's facial expressions, voice, heart rate, etc. and collects them as data.

[0285] Processing flow

[0286] Presentation of initial face image

[0287] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[0288] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[0289] User choice and sentiment data collection

[0290] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[0291] At the same time, the emotion engine recognizes the user's emotional data (happiness, surprise, interest, etc.) and transmits this information to the server via the device.

[0292] Learning preferences and generating new face images

[0293] The server receives the user's selection data and emotion data and trains an AI model based on the selected facial image features (e.g., eye shape, nose height, skin color, etc.) and emotion data to learn the user's preferences.

[0294] Based on the learning results, the server generates a new set of facial images that more closely match the user's preferences and sends them to the terminal.

[0295] The terminal again displays the new set of face images to the user, preparing for the next selection.

[0296] Iterative Process

[0297] The user again selects one of the newly presented facial images that they like best. This selection information and emotional data are also recorded on the device and sent to the server. At each step, the server repeatedly performs learning based on the user's selection information and emotional data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0298] Final facial image generation

[0299] After multiple selections and learning processes, the server finally generates a facial image that best reflects the user's preferences, taking into account emotional data, and sends this final facial image to the device.

[0300] Fine-tuning and checking

[0301] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[0302] This system allows users to avoid the hassle of inputting prompts and intuitively generate their preferred facial images with high accuracy. Furthermore, by using an emotion engine, the system can also reflect the user's instantaneous emotions, resulting in even more accurate results.

[0303] The processing flow will be explained below.

[0304] Step 1:

[0305] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[0306] Step 2:

[0307] The device displays the initial set of facial images received from the server to the user, arranging them in a layout that allows the user to easily review each image. At the same time, the emotion engine analyzes the user's facial expressions and prepares to collect initial emotion data.

[0308] Step 3:

[0309] The user selects the facial image they like best from the displayed facial images. Once the user has completed their selection, the device records the selected facial image information and sends it to the server. At the same time, the emotion engine recognizes the user's emotion at the time of selection (e.g., joy, surprise, interest, etc.), and this emotion data is also sent to the server via the device.

[0310] Step 4:

[0311] The server receives the user's selection data and emotion data, extracts features (e.g., eye shape, nose height, skin color, etc.) from the selected facial image, and trains an AI model based on the feature information and emotion data, thereby learning the user's preferences.

[0312] Step 5:

[0313] The server generates a new set of facial images that reflect the learned user's preferred features and emotional data. These facial images are adjusted to more closely match the user's preferences. The generated new set of facial images is then sent to the terminal.

[0314] Step 6:

[0315] The device will then display the newly sent face images to the user again, and the interface will be arranged so that the user can select a face image again, while the emotion engine continues to analyze the user's facial expressions and collect emotional data.

[0316] Step 7:

[0317] The user again selects one of the newly presented facial images that they like best. This selection information and emotion data are also recorded on the device and sent to the server. The server then learns again based on the user's selection information and emotion data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0318] Step 8:

[0319] After a predetermined number of selections and learning processes are completed, the server generates a final facial image that best reflects the user's preferences. This final facial image is generated by combining past selections and emotional data. This final facial image is then sent to the terminal.

[0320] Step 9:

[0321] The device displays the final facial image to the user. The user can review the facial image and make minor adjustments (e.g., changing the size of the eyes or adjusting the hairstyle) as needed. The details of the adjustments are also recorded on the device and, if necessary, sent to the server and reflected. The user can review the results of the adjustments and confirm the final facial image.

[0322] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input, learning preferences step by step, and based on real-time emotional data from the emotion engine, highly accurate results can be achieved.

[0323] Example 2

[0324] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0325] Conventional facial image generation systems require users to input specific prompts, making intuitive operation difficult. Furthermore, they are unable to reflect the user's instantaneous emotions, making it difficult to generate highly accurate facial images. Therefore, there is a demand for a system that can efficiently generate facial images that more accurately reflect the user's preferences.

[0326] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0327] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for recognizing and collecting user emotion data, means for learning the user's preferred features based on the collected selection data and emotion data, and means for generating a new facial image that reflects the learned features. This makes it possible to generate a highly accurate facial image that reflects the user's preferences and takes emotion data into consideration.

[0328] A "user" is a person who operates the system, selects a preferred facial image, and provides emotional data.

[0329] The "server" is a computer system that generates facial images using an image generation AI model and learns based on user selection information and emotional data.

[0330] A "terminal" is a device operated by a user (e.g., a PC or smartphone) that displays a facial image sent from a server and transmits the user's selection information and emotional data to the server.

[0331] An "image generation AI model" is an artificial intelligence model that learns a user's preferred features and generates facial images that reflect those features.

[0332] The "initial face images" are a plurality of face images that the system first presents to the user, and are images that are set to cover a wide range of user preferences.

[0333] "Selection information" is data generated when a user selects a preferred face image from among the presented face images.

[0334] "Emotion data" refers to data of emotions recognized from the user's facial expressions, voice, heart rate, etc.

[0335] An "emotion engine" is a system for recognizing user emotions and collecting emotional data.

[0336] "Learning" is the process by which the AI ​​model analyzes the user's preferred features based on the user's selection information and emotional data, and reflects this in future facial image generation.

[0337] The "new facial image" is a facial image that is closer to the user's preferences and is generated based on the user's selection information and emotion data.

[0338] "Fine tuning" is an operation that allows the user to adjust the details of the generated facial image.

[0339] This invention is a system for efficiently generating facial images that suit a user's preferences, and combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes: presenting multiple facial images, learning preferences based on the user's selection and emotions at the time, and generating new facial images.

[0340] Server Configuration

[0341] The server uses an image generation AI model (e.g., StyleGAN2) to generate facial images that reflect the user's preferences. The server generates an initial facial image and a new facial image based on the results of learning the user's preferences. The server also receives the user's selection information and emotional data and uses that information to learn.

[0342] Device configuration

[0343] The terminal is a device (e.g., PC or smartphone) operated by the user, and is responsible for displaying the facial image sent from the server to the user. It also transmits information about the facial image selected by the user and emotion data obtained from the emotion engine to the server.

[0344] User Roles

[0345] The user selects the facial image they like best from the displayed set of facial images and provides feedback to the system. The emotion engine also collects the user's emotional data.

[0346] Emotion engine configuration

[0347] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc. This data, along with the user's selection, is sent via the device to a server to help improve the accuracy of facial image generation.

[0348] Specific examples

[0349] 1. Generate and present an initial face image:

[0350] The server inputs the prompt "Generate 10 types of facial images" into the image generation AI model, and generates 10 types of initial facial images. These images are sent to the device in JPEG format.

[0351] The terminal displays the transmitted image to the user and provides an interface with selection buttons.

[0352] 2. Collecting user choice and sentiment data:

[0353] The user selects the face image they like best from the ten displayed face images. The selection information is recorded on the device and sent to the server in JSON format.

[0354] The emotion engine uses a camera and microphone to collect the user's facial expressions and tone of voice, as well as heart rate data from a heart rate sensor, which are also sent to the server.

[0355] 3. Generate a new face image:

[0356] The server trains an AI model based on the received selection data and emotion data, inputs the prompt "Please generate a new facial image based on the user's preferences," and generates a new facial image.

[0357] The new facial image is again sent to the terminal in JPEG format and displayed to the user.

[0358] 4. Repeat process:

[0359] The user then selects the face image they like best from the newly presented face images, collects the selection information and emotion data in the same way, and sends them to the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0360] 5. Generate the final face image:

[0361] After multiple selections and learning, the server generates a final facial image that best reflects the user's preferences and sends it to the terminal.

[0362] The user can make fine adjustments to the generated facial image, and the results are also recorded on the device and sent to the server if necessary.

[0363] Hardware and software used

[0364] Server: A high-performance computer system (e.g., a server machine equipped with a GPU)

[0365] Device: PC or smartphone

[0366] Image generation AI model: StyleGAN2

[0367] Prompt Sentence Examples

[0368] "Generate 10 different facial images."

[0369] "Generate new facial images based on user preferences."

[0370] "Generate a facial image based on the user's final preferences."

[0371] This system allows users to intuitively operate the system to accurately generate facial images of their choice. In addition, by using an emotion engine, the system can also reflect the user's momentary emotions, resulting in even more accurate results.

[0372] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0373] Step 1:

[0374] The server generates initial face images using an image generation AI model (e.g., StyleGAN2). The server inputs a prompt statement, "Please generate 10 types of face images," into the generative AI model. 10 types of initial face images (JPEG format) are generated as output.

[0375] Step 2:

[0376] The server sends the generated initial face image to the terminal, and as an output, the face image data is transferred to the terminal.

[0377] Step 3:

[0378] The terminal displays the received initial set of face images to the user, arranges the layout, and provides an interface that adds a selection button to each image. As an output, a set of face images is created that is displayed to the user.

[0379] Step 4:

[0380] The user selects one of the displayed face images that he or she likes the most. The face image that the user likes is selected as an input.

[0381] Step 5:

[0382] The device records the user's selection information and sends it to the server. The selection information is sent to the server in JSON format. As an output, the user's selection information is saved on the server.

[0383] Step 6:

[0384] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc., and sends this information to the server via the device. Emotional data is collected as input, and emotional data is stored on the server as output.

[0385] Step 7:

[0386] The server receives the user's selection data and emotion data and stores them in a database. The selection data and emotion data are sent to the server as input, and these data are recorded in the database as output.

[0387] Step 8:

[0388] The server trains an AI model based on the features (e.g., eye shape, nose height, skin color, etc.) and emotion data of the selected facial images. The feature data and emotion data are used as input, and the AI ​​model is updated as output.

[0389] Step 9:

[0390] The server inputs the prompt "Please generate new facial images based on the user's preferences" into the generative AI model and generates a set of new facial images. The new facial images (JPEG format) are generated as output.

[0391] Step 10:

[0392] The server sends the newly generated face images to the terminal. As an output, the face image data is transferred to the terminal.

[0393] Step 11:

[0394] The terminal displays the new set of face images to the user again and prepares for the next selection. The terminal provides an interface that adjusts the layout and adds a selection button again. As an output, a new set of face images is created and displayed to the user.

[0395] Step 12:

[0396] The user then selects the face image they like best from the newly presented set of face images. The preferred face image is selected as input.

[0397] Step 13:

[0398] The device again records the user's selection information and sends it to the server in JSON format. As an output, the user's selection information is saved on the server.

[0399] Step 14:

[0400] The emotion engine again recognizes the user's emotion data and sends this information to the server via the device. Emotion data is collected as input, and stored as output on the server.

[0401] Step 15:

[0402] The server then trains the AI ​​model again based on the user's selection and emotion data. This process is repeated a set number of times (e.g., 5-10 times), using feature data and emotion data as input and further refining the AI ​​model as output.

[0403] Step 16:

[0404] After multiple selections and learning are completed, the server inputs the prompt sentence "Please generate a facial image based on the user's final preferences" into the generative AI model to generate the final facial image. The final facial image (JPEG format) is generated as the output.

[0405] Step 17:

[0406] The server sends the final generated facial image to the terminal. As an output, the final facial image data is transferred to the terminal.

[0407] Step 18:

[0408] The terminal displays the final face image to the user. As an output, a final face image is created that is displayed to the user.

[0409] Step 19:

[0410] The user reviews this facial image and makes fine adjustments as needed, adding the fine adjustment data as input.

[0411] Step 20:

[0412] The terminal records the results of the fine-tuning and transmits them to the server for reflection as needed. As an output, the final face image is fine-tuned.

[0413] (Application example 2)

[0414] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0415] In conventional systems, simply generating a facial image that reflects a user's preferences takes a lot of time and effort, and they tend to ignore the user's emotions, making it difficult to generate optimal advertisements. Furthermore, advertisements are not sufficiently personalized based on facial image preferences, limiting their effectiveness.

[0416] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for presenting a plurality of facial images to the user, means for the user to select from the displayed facial images, and means for collecting data and emotional data based on the user's selection. This makes it possible to learn the user's preferred features based on the collected data and the user's emotional data, and to generate new facial images and personalized advertisements that reflect the learned features. This makes it possible to efficiently generate and distribute advertisements that reflect the user's preferences and emotions, maximizing advertising effectiveness.

[0417] A "system" is a collection of multiple components and processes that function together.

[0418] A "face image" is image data that digitally represents a person's face.

[0419] "Presenting" is the act of visually displaying information to a user.

[0420] A "means" is a device or method for achieving a specific function.

[0421] "Selection" is the act of choosing one option from multiple options.

[0422] "Data collection" is the process of obtaining information about user choices and responses.

[0423] "Emotional data" is digital information that indicates a user's emotional state.

[0424] "Learning" is the process by which artificial intelligence finds patterns based on data and improves its predictive capabilities.

[0425] "Generation" is the act of creating new digital content.

[0426] "Personalized advertising" refers to advertising content that is individually customized based on a user's preferences and characteristics.

[0427] "Fine-tuning" is the act of making minor modifications to the generated content.

[0428] This invention is a system that generates and delivers customized advertisements based on the user's preferences and emotions. The system collects data on the user's choices and emotions at the time, and learns the user's preferences based on this data. It also generates new facial images and personalized advertisements based on the learned data.

[0429] System configuration

[0430] server

[0431] It has an image generation AI model and generates facial images and advertisements.

[0432] It receives data from the emotion engine and uses it to learn user preferences.

[0433] An initial face image and a new face image and advertisement are generated to be displayed to the user.

[0434] Terminal

[0435] The device the user operates (e.g., PC, smartphone).

[0436] A facial image and advertisements are displayed to the user, and information and emotional data selected by the user are sent to a server.

[0437] user

[0438] Select your preferred facial image and provide feedback to the system.

[0439] Emotion data is collected through an emotion engine.

[0440] Hardware and software used

[0441] Smartphone (iOS or Android device)

[0442] Server (Cloud-based: AWS (registered trademark), Google (registered trademark), etc.)

[0443] Emotion engine (DeepFace, Emotion API, etc.)

[0444] Image generation AI models (GANs, StyleGAN, etc.)

[0445] Language models (GPT-4 (registered trademark), BERT, etc.)

[0446] Data processing and calculation

[0447] The server uses an image generation AI model to generate an initial set of facial images and sends them to the smartphone. The smartphone displays the received facial images to the user, allowing the user to select their preferred image. The selection information and data obtained from the emotion engine are then sent to the server. The server uses this data to train the AI ​​model and learn the user's preferences. The learning results are used to generate new facial images and personalized advertisements, which are then sent back to the smartphone.

[0448] Specific examples

[0449] For example, when a user selects their favorite facial image, the selection information is sent along with emotional data indicating "happiness." The server then uses this data to learn the user's preferences and generate the next facial image and advertisement. By repeating this process, highly accurate personalized advertisements are generated.

[0450] Example prompt sentence:

[0451] "Generate the most preferred face image based on the user's preferences. User preference data: {User preference data} Emotion data: {User emotion data}"

[0452] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0453] Step 1:

[0454] The server uses an image generation AI model to generate a set of initial face images (e.g., 10 different face images). These face images are designed to cover a wide range of user preferences. The generated face images are then sent from the server to the device.

[0455] Input: None (default setting)

[0456] Data processing / computation: Uses image generation AI models to generate facial images based on random or predefined patterns

[0457] Output: Initial face images

[0458] Specific behavior:

[0459] The server runs image generation AI models such as GANs and StyleGAN to generate facial images.

[0460] Step 2:

[0461] The device displays the initial set of facial images sent from the server to the user, and arranges the screen layout to make it easy for the user to select an image.

[0462] Input: Face images sent from the server

[0463] Data processing: Adjusting the layout of face images

[0464] Output: Display face image to user

[0465] Specific behavior:

[0466] Facial images are arranged on the device display and the user interface is designed.

[0467] Step 3:

[0468] The user selects the facial image they like best from the displayed images, and the device collects the user's emotional data (happiness, surprise, interest, etc.) through an emotion engine.

[0469] Input: User selects a face image

[0470] Data processing: Collecting selection information and emotion data

[0471] Output: Information and emotion data of selected face images

[0472] Specific behavior:

[0473] The user taps on the face image of their choice on the touchscreen, and at the same time the emotion engine measures emotional data using the camera and microphone.

[0474] Step 4:

[0475] The device sends user selection information and emotion data to the server, which uses the received data to train an AI model and learn user preferences.

[0476] Input: User selection information and emotion data

[0477] Data processing / computation: Training AI models

[0478] Output: Updated AI model

[0479] Specific behavior:

[0480] The server passes the features and emotional data of the selected facial images as input to the AI ​​model, which then uses machine learning algorithms to learn the user's preferences.

[0481] Step 5:

[0482] The server generates a new set of face images and a personalized advertisement based on the learning results, and transmits the generated face images and advertisement to the terminal again.

[0483] Input: Updated AI model

[0484] Data processing / computation: generating new facial images and advertisements

[0485] Output: New face images and personalized ads

[0486] Specific behavior:

[0487] The server then uses the image generation AI model again based on the learned preferred features to generate new facial images and advertisements.

[0488] Step 6:

[0489] The device displays the newly generated set of facial images and personalized advertisements to the user, and prepares the user to make another selection. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0490] Input: New face image and advertisement sent from the server

[0491] Data processing: Adjusting facial images and ad layout

[0492] Output: Redisplay to user

[0493] Specific behavior:

[0494] New facial images and advertisements are displayed on the device display, and the user interface is redesigned.

[0495] Step 7:

[0496] After the final selection, the server generates a final face image and advertisement that reflects the user's preferences to the maximum extent possible, and transmits them to the terminal.

[0497] Input: Multiple selection information and emotion data

[0498] Data processing / calculation: Generation of final face image and advertisement

[0499] Output: Final face image and advertisement

[0500] Specific behavior:

[0501] The server aggregates the data from all selections and generates the final facial image and personalized advertisements.

[0502] Step 8:

[0503] The device displays the final facial image and advertisement to the user, providing the user with the ability to make fine adjustments as needed.

[0504] Input: Final face image and advertisement

[0505] Data processing: Providing a fine-tuning interface

[0506] Output: Fine-tuned face image and advertisement

[0507] Specific behavior:

[0508] The final facial image and advertisement are displayed on the device display, and a user interface for fine-tuning is provided.

[0509] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0510] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0511] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0512] [Second embodiment]

[0513] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0514] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0515] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0516] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0517] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0518] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0519] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0520] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0521] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0522] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0523] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0524] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0525] The present invention provides a system for efficiently generating facial images that suit a user's preferences. The system includes a series of processes that present multiple facial images, learn the user's preferences based on the user's selection, and generate new facial images.

[0526] System configuration

[0527] server

[0528] It has an image generation AI model.

[0529] An initial face image and a new face image are generated.

[0530] It receives user selection information and learns based on that information.

[0531] Terminal

[0532] The device the user operates (e.g., PC, smartphone).

[0533] The facial image sent from the server is displayed to the user.

[0534] The information of the face image selected by the user is sent to the server.

[0535] user

[0536] Select your preferred facial image and provide feedback to the system.

[0537] Processing flow

[0538] Presentation of initial face image

[0539] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device.

[0540] The device displays these facial images to the user.

[0541] User Selection

[0542] The user selects the facial image that he or she likes best from the presented facial images.

[0543] The terminal transmits this selection information to the server.

[0544] Learning preferences

[0545] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[0546] Generating new face images

[0547] The server generates the next stage of facial images based on the learning results and sends them to the terminal again.

[0548] A new face image will be displayed that reflects the user's selection.

[0549] Iterative Process

[0550] The user again selects the face image they like best from the new set of face images, repeating this process multiple times to provide feedback to the system.

[0551] Through this process, the server gradually learns the user's preferences and generates highly accurate facial images.

[0552] Specific examples

[0553] 1. Presenting the initial face image:

[0554] A group of face images (e.g., A1, A2, A3, A4, A5) generated by the server are displayed on the terminal.

[0555] The user selects A3 and the device sends the information to the server.

[0556] 2. Learning preferences:

[0557] The server extracts features from A3's facial images and learns its preferences.

[0558] 3. Generate a new face image:

[0559] The server generates a new set of facial images (e.g., B1, B2, B3, B4, B5) based on the learning results and sends them to the terminal.

[0560] The user again selects B2, which they like best, and the process is repeated.

[0561] 4. Generate the final face image:

[0562] Finally, after multiple rounds of selection and learning, the server generates a facial image that best reflects the user's preferences and sends it to the device.

[0563] The user can review this facial image and make minor adjustments as needed.

[0564] This system allows users to avoid the hassle of prompt input and use intuitive operations to accurately generate their preferred facial images.

[0565] The processing flow will be explained below.

[0566] Step 1:

[0567] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[0568] Step 2:

[0569] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[0570] Step 3:

[0571] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[0572] Step 4:

[0573] The server receives the user's selection data, extracts features from the selected facial image (e.g., eye shape, nose height, skin color, etc.), and trains an AI model based on the feature information, thereby learning the user's preferences.

[0574] Step 5:

[0575] The server generates a new set of facial images that reflect the learned user's preferred features. These facial images are adjusted to more closely match the user's preferences. The newly generated set of facial images is then sent to the device.

[0576] Step 6:

[0577] The device will then display the newly sent face images to the user again, ready for the user to select the face image they like best.

[0578] Step 7:

[0579] The user again selects one of the newly presented face images that he or she likes best. This selection information is also recorded on the device and sent to the server. As in step 4, the user's selection information is again learned by the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0580] Step 8:

[0581] After multiple selections and learning processes are completed, the server finally generates a facial image that best reflects the user's preferences, and sends this final facial image to the device.

[0582] Step 9:

[0583] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[0584] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input and allowing the system to learn step-by-step based on user selection, highly accurate results can be achieved.

[0585] Example 1

[0586] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0587] Conventional facial image generation systems have struggled to efficiently generate the facial image desired by the user, often forcing the user to perform a huge amount of input work. Furthermore, due to a lack of systems that accurately respond to the user's preferences, the generated facial image often does not meet the user's expectations. The objective of this invention is to solve these problems and provide a system that enables users to intuitively operate the system and generate their preferred facial image with high accuracy.

[0588] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0589] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for transmitting the generated facial image to a user terminal, means for displaying the transmitted facial image to the user terminal, and means for the user to select a preferred facial image from the displayed new facial images. This allows the user to avoid the hassle of prompt input and to generate a preferred facial image with high accuracy through intuitive operation.

[0590] "User" refers to an individual who uses the System to select and generate their preferred facial image.

[0591] "Server" refers to a computer system equipped with an image generation AI model that generates and learns facial images based on user selection information.

[0592] "User terminal" refers to a device operated by a user (e.g., a PC or smartphone) that displays the facial image sent from the server and transmits the user's selection to the server.

[0593] "Image generation AI model" refers to an artificial intelligence model (e.g., StyleGAN) that generates facial images using deep learning technology.

[0594] "Facial Image" means an image of a human face that is generated by the system and presented to the user.

[0595] "Selection information" refers to data (such as the image ID and features) generated when a user selects one face image from multiple images.

[0596] "Features" refer to visual elements contained in a facial image (e.g., eye shape, nose height, skin color, etc.).

[0597] "Learning" refers to the process by which the AI ​​model analyzes and incorporates the user's preferences based on the information in the facial image selected by the user.

[0598] "Display" refers to the act of a user device providing a facial image to the user's visual sense.

[0599] "Generation" refers to the process by which an image-generating AI model generates new facial images.

[0600] "Fine-tuning" refers to the operation in which the user modifies the details of the final generated facial image to make it more closely match their own preferences.

[0601] This invention is a system for efficiently generating facial images that suit a user's preferences. This system consists of several main components, each of which performs a specific function. Specifically, it consists of a server, a user terminal, and a generative AI model.

[0602] System configuration

[0603] server

[0604] The server is equipped with an image generation AI model and has the following functions:

[0605] 1. Generate initial face images: Use an image generation AI model (e.g., StyleGAN) to generate a set of initial face images.

[0606] 2. Learning function: Based on the user's selection information, the system analyzes the characteristics of the facial image and learns the user's preferences.

[0607] 3. Generate new face images: The learned features are reflected to generate the next set of face images.

[0608] 4. Sending face image: Send the generated face image to the user terminal.

[0609] User Device

[0610] The user terminal is a device (e.g., PC, smartphone) that receives face images from the server and is operated by the user. It has the following functions:

[0611] 1. Displaying face image: The face image sent from the server is displayed to the user.

[0612] 2. Sending selected information: Send the information of the face image selected by the user to the server.

[0613] user

[0614] Users interact with the system as follows:

[0615] 1. Select a face image: Select the face image you like best from the displayed face images.

[0616] 2. Providing feedback: Giving feedback to the system through multiple selections.

[0617] Hardware and software used

[0618] Hardware: Servers equipped with high-performance GPUs, devices operated by users (PCs, smartphones)

[0619] Software: Generative AI model (e.g., StyleGAN), user interface (web application or mobile app)

[0620] Specific examples

[0621] 1. The server generates a set of initial face images using StyleGAN. For example, five face images (e.g., A1, A2, A3, A4, A5) are generated.

[0622] 2. The server sends the generated initial face images to the user's device.

[0623] 3. The user selects the facial image (e.g., A3) that they like best from the initial facial images displayed on the device. The device then sends this selection information to the server.

[0624] 4. The server analyzes the features of the selected face image (A3) and learns the user's preferences.

[0625] 5. The server generates a new set of face images (e.g., B1, B2, B3, B4, B5) reflecting the learning results and sends them again to the user's device.

[0626] 6. The user again selects the most preferred face image (e.g., B2) from the new set of face images, and the same process is repeated.

[0627] Prompt Sentence Examples

[0628] Design a system that allows a user to select their first preferred facial image and then learns its features to generate new facial images.

[0629] This system allows users to avoid the hassle of prompt input and to intuitively generate their preferred facial images with high accuracy.

[0630] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0631] Step 1:

[0632] The server generates a set of initial face images using an image generation AI model (e.g., StyleGAN), which uses a random latent space vector as input and generates multiple face images as output. Through this process, the system can obtain an initial set of face images (e.g., A1, A2, A3, A4, A5) to present to the user.

[0633] Step 2:

[0634] The server then sends the generated initial face images to the user's device. The generated face image data is used as input, and is sent to the user's device as output. This process enables the initial face images to be displayed on the device operated by the user.

[0635] Step 3:

[0636] The terminal displays the initial face images received from the server to the user. The received face image data is used as input and displayed on the user's screen as output. This process allows the user to check and select from the initial face images.

[0637] Step 4:

[0638] The user selects the face image they like best from the initial face images displayed. The visually displayed face image is used as input, and the ID and feature information of the selected face image are generated as output. This process determines the user's selection information.

[0639] Step 5:

[0640] The terminal sends the selection information of the face image selected by the user to the server. The user's selection information is used as input and sent to the server as output. This process allows the server to receive the user's preferences.

[0641] Step 6:

[0642] The server analyzes the features of the selected facial image based on the received selection information. The selected facial image data is used as input, and a feature vector is generated as output. This process provides the server with information to learn the user's preferred features.

[0643] Step 7:

[0644] The server inputs the analyzed feature vectors into a generative AI model to learn the user's preferences. The feature vectors are used as input, and updated model parameters are obtained as output. This process allows the generative AI model to reflect the user's preferences.

[0645] Step 8:

[0646] The server then generates a new set of facial images based on the learning results. The updated generative AI model is used as input, and new facial images (e.g., B1, B2, B3, B4, B5) are generated as output. This process provides the next phase of facial images to the user.

[0647] Step 9:

[0648] The server then sends the newly generated face images back to the user's device. The newly generated face image data is used as input, and is sent to the user's device as output. This process allows the user to check the new face images.

[0649] Step 10:

[0650] The device then presents a new set of face images to the user, using the new face image data sent from the server as input and displaying it on the user's screen as output. This process allows the user to select a face image again.

[0651] Step 11:

[0652] The user then selects the face image they like best from a set of new face images. The new face image displayed is used as input, and the information about the reselected face image is generated as output. This process is repeated multiple times until the face image closest to the user's preference is generated.

[0653] (Application example 1)

[0654] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0655] Conventional facial image generation systems have difficulty efficiently generating facial images that match a user's preferences. Furthermore, when applying the generated facial images to a virtual store clerk character, the user's preferences are not fully reflected. Therefore, improving the user experience has become a challenge.

[0656] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0657] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for presenting the generated facial images to the user in a virtual store and for the user to select one, and means for applying the newly generated facial image to a store clerk character in the virtual store. This allows the user to intuitively generate a preferred facial image and have it reflected in the store clerk character in the virtual store.

[0658] A "user" is a person who operates the system, selects his / her preferred facial image, and provides the system with the selected information.

[0659] A "face image" is an image with facial features that is presented for selection by the user.

[0660] The "presentation means" is a mechanism for displaying the facial image generated by the server on the terminal operated by the user.

[0661] The "means for making a selection" is a mechanism that provides an interface for a user to select a preferred facial image from a plurality of displayed facial images.

[0662] "Means for collecting data" refers to a mechanism for transmitting user selection information to a server and storing it.

[0663] The "means of learning features" is a mechanism that uses AI models to analyze and learn users' facial image preferences based on collected user selection data.

[0664] The "means for generating a new facial image" is a mechanism for generating a new facial image that reflects the user's preferences based on the learning results.

[0665] A "virtual store" is a store that exists on the Internet or in a virtual space and allows users to browse and purchase products online.

[0666] A "store clerk character" is a virtual character that appears in the virtual store to introduce and recommend products.

[0667] The present invention provides a system that allows users to efficiently generate facial images of their choice and apply the generated facial images to store clerk characters in a virtual store. This system operates mainly on a server and terminals and includes a process for learning and reflecting user preferences.

[0668] System configuration

[0669] server

[0670] It is equipped with an image generation AI model that generates initial and new facial images, and also receives user selection information and learns based on that information.

[0671] Terminal

[0672] A device operated by the user (e.g., PC, smartphone). It displays the facial images sent from the server to the user and sends the information of the facial image selected by the user to the server.

[0673] user

[0674] Select your preferred facial image and provide feedback to the system.

[0675] Processing flow

[0676] Presentation of initial face image

[0677] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device, where the user is shown these facial images.

[0678] User Selection

[0679] The user selects the facial image that they like best from the presented facial images, and the terminal transmits this selection information to the server.

[0680] Learning preferences

[0681] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[0682] Generating new face images

[0683] The server generates the next set of facial images based on the learning results and sends them back to the device, where new facial images reflecting the user's selection are displayed.

[0684] Applies to virtual stores

[0685] The generated facial image is applied to a store clerk character in a virtual store, and the user can interact with this store clerk character in the virtual store.

[0686] Hardware and software used

[0687] Hardware: Smartphone (iOS or Android), cloud server

[0688] Software: Python, PIL (Python Imaging Library), tkinter (for creating GUI)

[0689] Data processing and calculation

[0690] 1. Presentation of initial face images: Obtain the face image group URL generated by the server, download each image, and process it into display data.

[0691] 2. Sending user selection information: Display a dialog using tkinter, get the index of the image selected by the user, and send it to the server.

[0692] 3. Generate new face images: The server obtains the new face image URL and downloads each image.

[0693] 4. Iteration: The basic process is repeated until a facial image that satisfies the user is generated.

[0694] 5. Application in virtual stores: By applying the generated facial image to a clerk character in a virtual store, users can interact with their favorite clerk character.

[0695] Examples of concrete examples and prompts

[0696] Presentation of initial face image

[0697] The user selects the image they like best from the first five facial images displayed (e.g., A1, A2, A3, A4, A5).

[0698] Prompt: "Please choose the face image below that you like best."

[0699] Sending user selection information

[0700] For example, if the user selects A3, this information is sent to the server.

[0701] Prompt: "A3 selected. Generate a new face image."

[0702] Generating new face images

[0703] The server generates a set of new face images (e.g., B1, B2, B3, B4, B5) and presents them to the user again.

[0704] Prompt: "Facial images for the next step have been generated. Please choose the one you like best again."

[0705] In this way, the user can intuitively create a facial image of their choice and apply it to a salesperson character in a virtual store.

[0706] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0707] Step 1:

[0708] The server uses an image generation AI model to generate a set of initial facial images and sends their URLs to the device.

[0709] Input: A request comes to the server to generate the initial image.

[0710] Data processing / data calculation: The server generates initial facial images using an image generation AI model, generates a URL for each facial image, and sends it to the device.

[0711] Output: URL list of initial face images.

[0712] Step 2:

[0713] The device downloads the image using the URL of the initial face image received from the server and displays it to the user.

[0714] Input: URL list of initial face images sent from the server.

[0715] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[0716] Output: Multiple initial face images shown to the user.

[0717] Step 3:

[0718] The user selects the facial image that he or she likes best from the presented facial images.

[0719] Input: Multiple displayed face images.

[0720] Data processing / data calculation: Users intuitively select the facial image they like best.

[0721] Output: The index of the selected face image.

[0722] Step 4:

[0723] The terminal transmits the index of the face image selected by the user to the server.

[0724] Input: The index of the face image selected by the user.

[0725] Data processing / data calculation: Converts index information into data to be sent to the server.

[0726] Output: Sends the selection index to the server.

[0727] Step 5:

[0728] The server extracts features from the selected facial image and inputs them into an image generation AI model to learn the user's preferences.

[0729] Input: The index of the selected face image.

[0730] Data processing / data calculation: The server extracts the features of the selected facial image (eye shape, nose height, skin color, etc.) and uses an AI model to learn the user's preferences.

[0731] Output: The training results.

[0732] Step 6:

[0733] The server generates a new set of facial images based on the learning results and sends the URL to the terminal.

[0734] Input: Training results.

[0735] Data processing / data calculation: The server generates new facial images using an AI model and sends their URLs to the device.

[0736] Output: A list of URLs for new face images.

[0737] Step 7:

[0738] The device will use the new face image URL to download the image and display it to the user again.

[0739] Input: A list of URLs for new face images.

[0740] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[0741] Output: Multiple new face images that are displayed to the user.

[0742] Step 8:

[0743] Repeat steps 3 through 7 until the user is satisfied with the generation process.

[0744] Input: A new face image.

[0745] Data processing / data calculation: The user again selects the facial image they like best, and a new facial image is generated based on that.

[0746] Output: The final selected face image.

[0747] Step 9:

[0748] The generated final facial image is applied to a clerk character in a virtual store.

[0749] Input: The final selected face image.

[0750] Data processing / data calculation: The server converts the selected facial image into data to be applied to the virtual store clerk character.

[0751] Output: A custom store clerk character to be used in the virtual store.

[0752] Step 10:

[0753] Users interact with store clerk characters generated within the virtual store.

[0754] Input: A virtual store where the clerk character applies.

[0755] Data processing / data calculation: A store clerk character behaves within the virtual store and interacts with the user.

[0756] Output: A user experience interacting with a custom store clerk character.

[0757] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0758] This invention is a system for efficiently generating facial images that suit a user's preferences, and also combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes that present multiple facial images, learn the user's preferences based on the user's selection and the emotions at that time, and generate new facial images.

[0759] System configuration

[0760] server

[0761] It has an image generation AI model.

[0762] An initial face image and a new face image are generated.

[0763] It receives user selection information and emotion data and learns based on that information.

[0764] Terminal

[0765] The device the user operates (e.g., PC, smartphone).

[0766] The facial image sent from the server is displayed to the user.

[0767] Information about the face image selected by the user and emotion data obtained from the emotion engine are sent to the server.

[0768] user

[0769] Select your preferred facial image and provide feedback to the system.

[0770] Emotion data is collected by an emotion engine.

[0771] Emotion Engine

[0772] It recognizes emotions from the user's facial expressions, voice, heart rate, etc. and collects them as data.

[0773] Processing flow

[0774] Presentation of initial face image

[0775] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[0776] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[0777] User choice and sentiment data collection

[0778] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[0779] At the same time, the emotion engine recognizes the user's emotional data (happiness, surprise, interest, etc.) and transmits this information to the server via the device.

[0780] Learning preferences and generating new face images

[0781] The server receives the user's selection data and emotion data and trains an AI model based on the selected facial image features (e.g., eye shape, nose height, skin color, etc.) and emotion data to learn the user's preferences.

[0782] Based on the learning results, the server generates a new set of facial images that more closely match the user's preferences and sends them to the terminal.

[0783] The terminal again displays the new set of face images to the user, preparing for the next selection.

[0784] Iterative Process

[0785] The user again selects one of the newly presented facial images that they like best. This selection information and emotional data are also recorded on the device and sent to the server. At each step, the server repeatedly performs learning based on the user's selection information and emotional data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0786] Final facial image generation

[0787] After multiple selections and learning processes, the server finally generates a facial image that best reflects the user's preferences, taking into account emotional data, and sends this final facial image to the device.

[0788] Fine-tuning and checking

[0789] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[0790] This system allows users to avoid the hassle of inputting prompts and intuitively generate their preferred facial images with high accuracy. Furthermore, by using an emotion engine, the system can also reflect the user's instantaneous emotions, resulting in even more accurate results.

[0791] The processing flow will be explained below.

[0792] Step 1:

[0793] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[0794] Step 2:

[0795] The device displays the initial set of facial images received from the server to the user, arranging them in a layout that allows the user to easily review each image. At the same time, the emotion engine analyzes the user's facial expressions and prepares to collect initial emotion data.

[0796] Step 3:

[0797] The user selects the facial image they like best from the displayed facial images. Once the user has completed their selection, the device records the selected facial image information and sends it to the server. At the same time, the emotion engine recognizes the user's emotion at the time of selection (e.g., joy, surprise, interest, etc.), and this emotion data is also sent to the server via the device.

[0798] Step 4:

[0799] The server receives the user's selection data and emotion data, extracts features (e.g., eye shape, nose height, skin color, etc.) from the selected facial image, and trains an AI model based on the feature information and emotion data, thereby learning the user's preferences.

[0800] Step 5:

[0801] The server generates a new set of facial images that reflect the learned user's preferred features and emotional data. These facial images are adjusted to more closely match the user's preferences. The generated new set of facial images is then sent to the terminal.

[0802] Step 6:

[0803] The device will then display the newly sent face images to the user again, and the interface will be arranged so that the user can select a face image again, while the emotion engine continues to analyze the user's facial expressions and collect emotional data.

[0804] Step 7:

[0805] The user again selects one of the newly presented facial images that they like best. This selection information and emotion data are also recorded on the device and sent to the server. The server then learns again based on the user's selection information and emotion data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0806] Step 8:

[0807] After a predetermined number of selections and learning processes are completed, the server generates a final facial image that best reflects the user's preferences. This final facial image is generated by combining past selections and emotional data. This final facial image is then sent to the terminal.

[0808] Step 9:

[0809] The device displays the final facial image to the user. The user can review the facial image and make minor adjustments (e.g., changing the size of the eyes or adjusting the hairstyle) as needed. The details of the adjustments are also recorded on the device and, if necessary, sent to the server and reflected. The user can review the results of the adjustments and confirm the final facial image.

[0810] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input, learning preferences step by step, and based on real-time emotional data from the emotion engine, highly accurate results can be achieved.

[0811] Example 2

[0812] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0813] Conventional facial image generation systems require users to input specific prompts, making intuitive operation difficult. Furthermore, they are unable to reflect the user's instantaneous emotions, making it difficult to generate highly accurate facial images. Therefore, there is a demand for a system that can efficiently generate facial images that more accurately reflect the user's preferences.

[0814] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0815] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for recognizing and collecting user emotion data, means for learning the user's preferred features based on the collected selection data and emotion data, and means for generating a new facial image that reflects the learned features. This makes it possible to generate a highly accurate facial image that reflects the user's preferences and takes emotion data into consideration.

[0816] A "user" is a person who operates the system, selects a preferred facial image, and provides emotional data.

[0817] The "server" is a computer system that generates facial images using an image generation AI model and learns based on user selection information and emotional data.

[0818] A "terminal" is a device operated by a user (e.g., a PC or smartphone) that displays a facial image sent from a server and transmits the user's selection information and emotional data to the server.

[0819] An "image generation AI model" is an artificial intelligence model that learns a user's preferred features and generates facial images that reflect those features.

[0820] The "initial face images" are a plurality of face images that the system first presents to the user, and are images that are set to cover a wide range of user preferences.

[0821] "Selection information" is data generated when a user selects a preferred face image from among the presented face images.

[0822] "Emotion data" refers to data of emotions recognized from the user's facial expressions, voice, heart rate, etc.

[0823] An "emotion engine" is a system for recognizing user emotions and collecting emotional data.

[0824] "Learning" is the process by which the AI ​​model analyzes the user's preferred features based on the user's selection information and emotional data, and reflects this in future facial image generation.

[0825] The "new facial image" is a facial image that is closer to the user's preferences and is generated based on the user's selection information and emotion data.

[0826] "Fine tuning" is an operation that allows the user to adjust the details of the generated facial image.

[0827] This invention is a system for efficiently generating facial images that suit a user's preferences, and combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes: presenting multiple facial images, learning preferences based on the user's selection and emotions at the time, and generating new facial images.

[0828] Server Configuration

[0829] The server uses an image generation AI model (e.g., StyleGAN2) to generate facial images that reflect the user's preferences. The server generates an initial facial image and a new facial image based on the results of learning the user's preferences. The server also receives the user's selection information and emotional data and uses that information to learn.

[0830] Device configuration

[0831] The terminal is a device (e.g., PC or smartphone) operated by the user, and is responsible for displaying the facial image sent from the server to the user. It also transmits information about the facial image selected by the user and emotion data obtained from the emotion engine to the server.

[0832] User Roles

[0833] The user selects the facial image they like best from the displayed set of facial images and provides feedback to the system. The emotion engine also collects the user's emotional data.

[0834] Emotion engine configuration

[0835] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc. This data, along with the user's selection, is sent via the device to a server to help improve the accuracy of facial image generation.

[0836] Specific examples

[0837] 1. Generate and present an initial face image:

[0838] The server inputs the prompt "Generate 10 types of facial images" into the image generation AI model, and generates 10 types of initial facial images. These images are sent to the device in JPEG format.

[0839] The terminal displays the transmitted image to the user and provides an interface with selection buttons.

[0840] 2. Collecting user choice and sentiment data:

[0841] The user selects the face image they like best from the ten displayed face images. The selection information is recorded on the device and sent to the server in JSON format.

[0842] The emotion engine uses a camera and microphone to collect the user's facial expressions and tone of voice, as well as heart rate data from a heart rate sensor, which are also sent to the server.

[0843] 3. Generate a new face image:

[0844] The server trains an AI model based on the received selection data and emotion data, inputs the prompt "Please generate a new facial image based on the user's preferences," and generates a new facial image.

[0845] The new facial image is again sent to the terminal in JPEG format and displayed to the user.

[0846] 4. Repeat process:

[0847] The user then selects the face image they like best from the newly presented face images, collects the selection information and emotion data in the same way, and sends them to the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0848] 5. Generate the final face image:

[0849] After multiple selections and learning, the server generates a final facial image that best reflects the user's preferences and sends it to the terminal.

[0850] The user can make fine adjustments to the generated facial image, and the results are also recorded on the device and sent to the server if necessary.

[0851] Hardware and software used

[0852] Server: A high-performance computer system (e.g., a server machine equipped with a GPU)

[0853] Device: PC or smartphone

[0854] Image generation AI model: StyleGAN2

[0855] Prompt Sentence Examples

[0856] "Generate 10 different facial images."

[0857] "Generate new facial images based on user preferences."

[0858] "Generate a facial image based on the user's final preferences."

[0859] This system allows users to intuitively operate the system to accurately generate facial images of their choice. In addition, by using an emotion engine, the system can also reflect the user's momentary emotions, resulting in even more accurate results.

[0860] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0861] Step 1:

[0862] The server generates initial face images using an image generation AI model (e.g., StyleGAN2). The server inputs a prompt statement, "Please generate 10 types of face images," into the generative AI model. 10 types of initial face images (JPEG format) are generated as output.

[0863] Step 2:

[0864] The server sends the generated initial face image to the terminal, and as an output, the face image data is transferred to the terminal.

[0865] Step 3:

[0866] The terminal displays the received initial set of face images to the user, arranges the layout, and provides an interface that adds a selection button to each image. As an output, a set of face images is created that is displayed to the user.

[0867] Step 4:

[0868] The user selects one of the displayed face images that he or she likes the most. The face image that the user likes is selected as an input.

[0869] Step 5:

[0870] The device records the user's selection information and sends it to the server. The selection information is sent to the server in JSON format. As an output, the user's selection information is saved on the server.

[0871] Step 6:

[0872] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc., and sends this information to the server via the device. Emotional data is collected as input, and emotional data is stored on the server as output.

[0873] Step 7:

[0874] The server receives the user's selection data and emotion data and stores them in a database. The selection data and emotion data are sent to the server as input, and these data are recorded in the database as output.

[0875] Step 8:

[0876] The server trains an AI model based on the features (e.g., eye shape, nose height, skin color, etc.) and emotion data of the selected facial images. The feature data and emotion data are used as input, and the AI ​​model is updated as output.

[0877] Step 9:

[0878] The server inputs the prompt "Please generate new facial images based on the user's preferences" into the generative AI model and generates a set of new facial images. The new facial images (JPEG format) are generated as output.

[0879] Step 10:

[0880] The server sends the newly generated face images to the terminal. As an output, the face image data is transferred to the terminal.

[0881] Step 11:

[0882] The terminal displays the new set of face images to the user again and prepares for the next selection. The terminal provides an interface that adjusts the layout and adds a selection button again. As an output, a new set of face images is created and displayed to the user.

[0883] Step 12:

[0884] The user then selects the face image they like best from the newly presented set of face images. The preferred face image is selected as input.

[0885] Step 13:

[0886] The device again records the user's selection information and sends it to the server in JSON format. As an output, the user's selection information is saved on the server.

[0887] Step 14:

[0888] The emotion engine again recognizes the user's emotion data and sends this information to the server via the device. Emotion data is collected as input, and stored as output on the server.

[0889] Step 15:

[0890] The server then trains the AI ​​model again based on the user's selection and emotion data. This process is repeated a set number of times (e.g., 5-10 times), using feature data and emotion data as input and further refining the AI ​​model as output.

[0891] Step 16:

[0892] After multiple selections and learning are completed, the server inputs the prompt sentence "Please generate a facial image based on the user's final preferences" into the generative AI model to generate the final facial image. The final facial image (JPEG format) is generated as the output.

[0893] Step 17:

[0894] The server sends the final generated facial image to the terminal. As an output, the final facial image data is transferred to the terminal.

[0895] Step 18:

[0896] The terminal displays the final face image to the user. As an output, a final face image is created that is displayed to the user.

[0897] Step 19:

[0898] The user reviews this facial image and makes fine adjustments as needed, adding the fine adjustment data as input.

[0899] Step 20:

[0900] The terminal records the results of the fine-tuning and transmits them to the server for reflection as needed. As an output, the final face image is fine-tuned.

[0901] (Application example 2)

[0902] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0903] In conventional systems, simply generating a facial image that reflects a user's preferences takes a lot of time and effort, and they tend to ignore the user's emotions, making it difficult to generate optimal advertisements. Furthermore, advertisements are not sufficiently personalized based on facial image preferences, limiting their effectiveness.

[0904] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for presenting a plurality of facial images to the user, means for the user to select from the displayed facial images, and means for collecting data and emotional data based on the user's selection. This makes it possible to learn the user's preferred features based on the collected data and the user's emotional data, and to generate new facial images and personalized advertisements that reflect the learned features. This makes it possible to efficiently generate and distribute advertisements that reflect the user's preferences and emotions, maximizing advertising effectiveness.

[0905] A "system" is a collection of multiple components and processes that function together.

[0906] A "face image" is image data that digitally represents a person's face.

[0907] "Presenting" is the act of visually displaying information to a user.

[0908] A "means" is a device or method for achieving a specific function.

[0909] "Selection" is the act of choosing one option from multiple options.

[0910] "Data collection" is the process of obtaining information about user choices and responses.

[0911] "Emotional data" is digital information that indicates a user's emotional state.

[0912] "Learning" is the process by which artificial intelligence finds patterns based on data and improves its predictive capabilities.

[0913] "Generation" is the act of creating new digital content.

[0914] "Personalized advertising" refers to advertising content that is individually customized based on a user's preferences and characteristics.

[0915] "Fine-tuning" is the act of making minor modifications to the generated content.

[0916] This invention is a system that generates and delivers customized advertisements based on the user's preferences and emotions. The system collects data on the user's choices and emotions at the time, and learns the user's preferences based on this data. It also generates new facial images and personalized advertisements based on the learned data.

[0917] System configuration

[0918] server

[0919] It has an image generation AI model and generates facial images and advertisements.

[0920] It receives data from the emotion engine and uses it to learn user preferences.

[0921] An initial face image and a new face image and advertisement are generated to be displayed to the user.

[0922] Terminal

[0923] The device the user operates (e.g., PC, smartphone).

[0924] A facial image and advertisements are displayed to the user, and information and emotional data selected by the user are sent to a server.

[0925] user

[0926] Select your preferred facial image and provide feedback to the system.

[0927] Emotion data is collected through an emotion engine.

[0928] Hardware and software used

[0929] Smartphone (iOS or Android device)

[0930] Server (Cloud-based: AWS, Google Cloud Platform, etc.)

[0931] Emotion engine (DeepFace, Emotion API, etc.)

[0932] Image generation AI models (GANs, StyleGAN, etc.)

[0933] Language models (GPT-4, BERT, etc.)

[0934] Data processing and calculation

[0935] The server uses an image generation AI model to generate an initial set of facial images and sends them to the smartphone. The smartphone displays the received facial images to the user, allowing the user to select their preferred image. The selection information and data obtained from the emotion engine are then sent to the server. The server uses this data to train the AI ​​model and learn the user's preferences. The learning results are used to generate new facial images and personalized advertisements, which are then sent back to the smartphone.

[0936] Specific examples

[0937] For example, when a user selects their favorite facial image, the selection information is sent along with emotional data indicating "happiness." The server then uses this data to learn the user's preferences and generate the next facial image and advertisement. By repeating this process, highly accurate personalized advertisements are generated.

[0938] Example prompt sentence:

[0939] "Generate the most preferred face image based on the user's preferences. User preference data: {User preference data} Emotion data: {User emotion data}"

[0940] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0941] Step 1:

[0942] The server uses an image generation AI model to generate a set of initial face images (e.g., 10 different face images). These face images are designed to cover a wide range of user preferences. The generated face images are then sent from the server to the device.

[0943] Input: None (default setting)

[0944] Data processing / computation: Uses image generation AI models to generate facial images based on random or predefined patterns

[0945] Output: Initial face images

[0946] Specific behavior:

[0947] The server runs image generation AI models such as GANs and StyleGAN to generate facial images.

[0948] Step 2:

[0949] The device displays the initial set of facial images sent from the server to the user, and arranges the screen layout to make it easy for the user to select an image.

[0950] Input: Face images sent from the server

[0951] Data processing: Adjusting the layout of face images

[0952] Output: Display face image to user

[0953] Specific behavior:

[0954] Facial images are arranged on the device display and the user interface is designed.

[0955] Step 3:

[0956] The user selects the facial image they like best from the displayed images, and the device collects the user's emotional data (happiness, surprise, interest, etc.) through an emotion engine.

[0957] Input: User selects a face image

[0958] Data processing: Collecting selection information and emotion data

[0959] Output: Information and emotion data of selected face images

[0960] Specific behavior:

[0961] The user taps on the face image of their choice on the touchscreen, and at the same time the emotion engine measures emotional data using the camera and microphone.

[0962] Step 4:

[0963] The device sends user selection information and emotion data to the server, which uses the received data to train an AI model and learn user preferences.

[0964] Input: User selection information and emotion data

[0965] Data processing / computation: Training AI models

[0966] Output: Updated AI model

[0967] Specific behavior:

[0968] The server passes the features and emotional data of the selected facial images as input to the AI ​​model, which then uses machine learning algorithms to learn the user's preferences.

[0969] Step 5:

[0970] The server generates a new set of face images and a personalized advertisement based on the learning results, and transmits the generated face images and advertisement to the terminal again.

[0971] Input: Updated AI model

[0972] Data processing / computation: generating new facial images and advertisements

[0973] Output: New face images and personalized ads

[0974] Specific behavior:

[0975] The server then uses the image generation AI model again based on the learned preferred features to generate new facial images and advertisements.

[0976] Step 6:

[0977] The device displays the newly generated set of facial images and personalized advertisements to the user, and prepares the user to make another selection. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[0978] Input: New face image and advertisement sent from the server

[0979] Data processing: Adjusting facial images and ad layout

[0980] Output: Redisplay to user

[0981] Specific behavior:

[0982] New facial images and advertisements are displayed on the device display, and the user interface is redesigned.

[0983] Step 7:

[0984] After the final selection, the server generates a final face image and advertisement that reflects the user's preferences to the maximum extent possible, and transmits them to the terminal.

[0985] Input: Multiple selection information and emotion data

[0986] Data processing / calculation: Generation of final face image and advertisement

[0987] Output: Final face image and advertisement

[0988] Specific behavior:

[0989] The server aggregates the data from all selections and generates the final facial image and personalized advertisements.

[0990] Step 8:

[0991] The device displays the final facial image and advertisement to the user, providing the user with the ability to make fine adjustments as needed.

[0992] Input: Final face image and advertisement

[0993] Data processing: Providing a fine-tuning interface

[0994] Output: Fine-tuned face image and advertisement

[0995] Specific behavior:

[0996] The final facial image and advertisement are displayed on the device display, and a user interface for fine-tuning is provided.

[0997] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0998] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0999] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1000] [Third embodiment]

[1001] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1002] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1003] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1004] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1005] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1006] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1007] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1008] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1009] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1010] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1011] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1012] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[1013] The present invention provides a system for efficiently generating facial images that suit a user's preferences. The system includes a series of processes that present multiple facial images, learn the user's preferences based on the user's selection, and generate new facial images.

[1014] System configuration

[1015] server

[1016] It has an image generation AI model.

[1017] An initial face image and a new face image are generated.

[1018] It receives user selection information and learns based on that information.

[1019] Terminal

[1020] The device the user operates (e.g., PC, smartphone).

[1021] The facial image sent from the server is displayed to the user.

[1022] The information of the face image selected by the user is sent to the server.

[1023] user

[1024] Select your preferred facial image and provide feedback to the system.

[1025] Processing flow

[1026] Presentation of initial face image

[1027] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device.

[1028] The device displays these facial images to the user.

[1029] User Selection

[1030] The user selects the facial image that he or she likes best from the presented facial images.

[1031] The terminal transmits this selection information to the server.

[1032] Learning preferences

[1033] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[1034] Generating new face images

[1035] The server generates the next stage of facial images based on the learning results and sends them to the terminal again.

[1036] A new face image will be displayed that reflects the user's selection.

[1037] Iterative Process

[1038] The user again selects the face image they like best from the new set of face images, repeating this process multiple times to provide feedback to the system.

[1039] Through this process, the server gradually learns the user's preferences and generates highly accurate facial images.

[1040] Specific examples

[1041] 1. Presenting the initial face image:

[1042] A group of face images (e.g., A1, A2, A3, A4, A5) generated by the server are displayed on the terminal.

[1043] The user selects A3 and the device sends the information to the server.

[1044] 2. Learning preferences:

[1045] The server extracts features from A3's facial images and learns its preferences.

[1046] 3. Generate a new face image:

[1047] The server generates a new set of facial images (e.g., B1, B2, B3, B4, B5) based on the learning results and sends them to the terminal.

[1048] The user again selects B2, which they like best, and the process is repeated.

[1049] 4. Generate the final face image:

[1050] Finally, after multiple rounds of selection and learning, the server generates a facial image that best reflects the user's preferences and sends it to the device.

[1051] The user can review this facial image and make minor adjustments as needed.

[1052] This system allows users to avoid the hassle of prompt input and use intuitive operations to accurately generate their preferred facial images.

[1053] The processing flow will be explained below.

[1054] Step 1:

[1055] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[1056] Step 2:

[1057] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[1058] Step 3:

[1059] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[1060] Step 4:

[1061] The server receives the user's selection data, extracts features from the selected facial image (e.g., eye shape, nose height, skin color, etc.), and trains an AI model based on the feature information, thereby learning the user's preferences.

[1062] Step 5:

[1063] The server generates a new set of facial images that reflect the learned user's preferred features. These facial images are adjusted to more closely match the user's preferences. The newly generated set of facial images is then sent to the device.

[1064] Step 6:

[1065] The device will then display the newly sent face images to the user again, ready for the user to select the face image they like best.

[1066] Step 7:

[1067] The user again selects one of the newly presented face images that he or she likes best. This selection information is also recorded on the device and sent to the server. As in step 4, the user's selection information is again learned by the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1068] Step 8:

[1069] After multiple selections and learning processes are completed, the server finally generates a facial image that best reflects the user's preferences, and sends this final facial image to the device.

[1070] Step 9:

[1071] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[1072] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input and allowing the system to learn step-by-step based on user selection, highly accurate results can be achieved.

[1073] Example 1

[1074] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1075] Conventional facial image generation systems have struggled to efficiently generate the facial image desired by the user, often forcing the user to perform a huge amount of input work. Furthermore, due to a lack of systems that accurately respond to the user's preferences, the generated facial image often does not meet the user's expectations. The objective of this invention is to solve these problems and provide a system that enables users to intuitively operate the system and generate their preferred facial image with high accuracy.

[1076] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1077] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for transmitting the generated facial image to a user terminal, means for displaying the transmitted facial image to the user terminal, and means for the user to select a preferred facial image from the displayed new facial images. This allows the user to avoid the hassle of prompt input and to generate a preferred facial image with high accuracy through intuitive operation.

[1078] "User" refers to an individual who uses the System to select and generate their preferred facial image.

[1079] "Server" refers to a computer system equipped with an image generation AI model that generates and learns facial images based on user selection information.

[1080] "User terminal" refers to a device operated by a user (e.g., a PC or smartphone) that displays the facial image sent from the server and transmits the user's selection to the server.

[1081] "Image generation AI model" refers to an artificial intelligence model (e.g., StyleGAN) that generates facial images using deep learning technology.

[1082] "Facial Image" means an image of a human face that is generated by the system and presented to the user.

[1083] "Selection information" refers to data (such as the image ID and features) generated when a user selects one face image from multiple images.

[1084] "Features" refer to visual elements contained in a facial image (e.g., eye shape, nose height, skin color, etc.).

[1085] "Learning" refers to the process by which the AI ​​model analyzes and incorporates the user's preferences based on the information in the facial image selected by the user.

[1086] "Display" refers to the act of a user device providing a facial image to the user's visual sense.

[1087] "Generation" refers to the process by which an image-generating AI model generates new facial images.

[1088] "Fine-tuning" refers to the operation in which the user modifies the details of the final generated facial image to make it more closely match their own preferences.

[1089] This invention is a system for efficiently generating facial images that suit a user's preferences. This system consists of several main components, each of which performs a specific function. Specifically, it consists of a server, a user terminal, and a generative AI model.

[1090] System configuration

[1091] server

[1092] The server is equipped with an image generation AI model and has the following functions:

[1093] 1. Generate initial face images: Use an image generation AI model (e.g., StyleGAN) to generate a set of initial face images.

[1094] 2. Learning function: Based on the user's selection information, the system analyzes the characteristics of the facial image and learns the user's preferences.

[1095] 3. Generate new face images: The learned features are reflected to generate the next set of face images.

[1096] 4. Sending face image: Send the generated face image to the user terminal.

[1097] User Device

[1098] The user terminal is a device (e.g., PC, smartphone) that receives face images from the server and is operated by the user. It has the following functions:

[1099] 1. Displaying face image: The face image sent from the server is displayed to the user.

[1100] 2. Sending selected information: Send the information of the face image selected by the user to the server.

[1101] user

[1102] Users interact with the system as follows:

[1103] 1. Select a face image: Select the face image you like best from the displayed face images.

[1104] 2. Providing feedback: Giving feedback to the system through multiple selections.

[1105] Hardware and software used

[1106] Hardware: Servers equipped with high-performance GPUs, devices operated by users (PCs, smartphones)

[1107] Software: Generative AI model (e.g., StyleGAN), user interface (web application or mobile app)

[1108] Specific examples

[1109] 1. The server generates a set of initial face images using StyleGAN. For example, five face images (e.g., A1, A2, A3, A4, A5) are generated.

[1110] 2. The server sends the generated initial face images to the user's device.

[1111] 3. The user selects the facial image (e.g., A3) that they like best from the initial facial images displayed on the device. The device then sends this selection information to the server.

[1112] 4. The server analyzes the features of the selected face image (A3) and learns the user's preferences.

[1113] 5. The server generates a new set of face images (e.g., B1, B2, B3, B4, B5) reflecting the learning results and sends them again to the user's device.

[1114] 6. The user again selects the most preferred face image (e.g., B2) from the new set of face images, and the same process is repeated.

[1115] Prompt Sentence Examples

[1116] Design a system that allows a user to select their first preferred facial image and then learns its features to generate new facial images.

[1117] This system allows users to avoid the hassle of prompt input and to intuitively generate their preferred facial images with high accuracy.

[1118] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1119] Step 1:

[1120] The server generates a set of initial face images using an image generation AI model (e.g., StyleGAN), which uses a random latent space vector as input and generates multiple face images as output. Through this process, the system can obtain an initial set of face images (e.g., A1, A2, A3, A4, A5) to present to the user.

[1121] Step 2:

[1122] The server then sends the generated initial face images to the user's device. The generated face image data is used as input, and is sent to the user's device as output. This process enables the initial face images to be displayed on the device operated by the user.

[1123] Step 3:

[1124] The terminal displays the initial face images received from the server to the user. The received face image data is used as input and displayed on the user's screen as output. This process allows the user to check and select from the initial face images.

[1125] Step 4:

[1126] The user selects the face image they like best from the initial face images displayed. The visually displayed face image is used as input, and the ID and feature information of the selected face image are generated as output. This process determines the user's selection information.

[1127] Step 5:

[1128] The terminal sends the selection information of the face image selected by the user to the server. The user's selection information is used as input and sent to the server as output. This process allows the server to receive the user's preferences.

[1129] Step 6:

[1130] The server analyzes the features of the selected facial image based on the received selection information. The selected facial image data is used as input, and a feature vector is generated as output. This process provides the server with information to learn the user's preferred features.

[1131] Step 7:

[1132] The server inputs the analyzed feature vectors into a generative AI model to learn the user's preferences. The feature vectors are used as input, and updated model parameters are obtained as output. This process allows the generative AI model to reflect the user's preferences.

[1133] Step 8:

[1134] The server then generates a new set of facial images based on the learning results. The updated generative AI model is used as input, and new facial images (e.g., B1, B2, B3, B4, B5) are generated as output. This process provides the next phase of facial images to the user.

[1135] Step 9:

[1136] The server then sends the newly generated face images back to the user's device. The newly generated face image data is used as input, and is sent to the user's device as output. This process allows the user to check the new face images.

[1137] Step 10:

[1138] The device then presents a new set of face images to the user, using the new face image data sent from the server as input and displaying it on the user's screen as output. This process allows the user to select a face image again.

[1139] Step 11:

[1140] The user then selects the face image they like best from a set of new face images. The new face image displayed is used as input, and the information about the reselected face image is generated as output. This process is repeated multiple times until the face image closest to the user's preference is generated.

[1141] (Application example 1)

[1142] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1143] Conventional facial image generation systems have difficulty efficiently generating facial images that match a user's preferences. Furthermore, when applying the generated facial images to a virtual store clerk character, the user's preferences are not fully reflected. Therefore, improving the user experience has become a challenge.

[1144] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1145] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for presenting the generated facial images to the user in a virtual store and for the user to select one, and means for applying the newly generated facial image to a store clerk character in the virtual store. This allows the user to intuitively generate a preferred facial image and have it reflected in the store clerk character in the virtual store.

[1146] A "user" is a person who operates the system, selects his / her preferred facial image, and provides the system with the selected information.

[1147] A "face image" is an image with facial features that is presented for selection by the user.

[1148] The "presentation means" is a mechanism for displaying the facial image generated by the server on the terminal operated by the user.

[1149] The "means for making a selection" is a mechanism that provides an interface for a user to select a preferred facial image from a plurality of displayed facial images.

[1150] "Means for collecting data" refers to a mechanism for transmitting user selection information to a server and storing it.

[1151] The "means of learning features" is a mechanism that uses AI models to analyze and learn users' facial image preferences based on collected user selection data.

[1152] The "means for generating a new facial image" is a mechanism for generating a new facial image that reflects the user's preferences based on the learning results.

[1153] A "virtual store" is a store that exists on the Internet or in a virtual space and allows users to browse and purchase products online.

[1154] A "store clerk character" is a virtual character that appears in the virtual store to introduce and recommend products.

[1155] The present invention provides a system that allows users to efficiently generate facial images of their choice and apply the generated facial images to store clerk characters in a virtual store. This system operates mainly on a server and terminals and includes a process for learning and reflecting user preferences.

[1156] System configuration

[1157] server

[1158] It is equipped with an image generation AI model that generates initial and new facial images, and also receives user selection information and learns based on that information.

[1159] Terminal

[1160] A device operated by the user (e.g., PC, smartphone). It displays the facial images sent from the server to the user and sends the information of the facial image selected by the user to the server.

[1161] user

[1162] Select your preferred facial image and provide feedback to the system.

[1163] Processing flow

[1164] Presentation of initial face image

[1165] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device, where the user is shown these facial images.

[1166] User Selection

[1167] The user selects the facial image that they like best from the presented facial images, and the terminal transmits this selection information to the server.

[1168] Learning preferences

[1169] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[1170] Generating new face images

[1171] The server generates the next set of facial images based on the learning results and sends them back to the device, where new facial images reflecting the user's selection are displayed.

[1172] Applies to virtual stores

[1173] The generated facial image is applied to a store clerk character in a virtual store, and the user can interact with this store clerk character in the virtual store.

[1174] Hardware and software used

[1175] Hardware: Smartphone (iOS or Android), cloud server

[1176] Software: Python, PIL (Python Imaging Library), tkinter (for creating GUI)

[1177] Data processing and calculation

[1178] 1. Presentation of initial face images: Obtain the face image group URL generated by the server, download each image, and process it into display data.

[1179] 2. Sending user selection information: Display a dialog using tkinter, get the index of the image selected by the user, and send it to the server.

[1180] 3. Generate new face images: The server obtains the new face image URL and downloads each image.

[1181] 4. Iteration: The basic process is repeated until a facial image that satisfies the user is generated.

[1182] 5. Application in virtual stores: By applying the generated facial image to a clerk character in a virtual store, users can interact with their favorite clerk character.

[1183] Examples of concrete examples and prompts

[1184] Presentation of initial face image

[1185] The user selects the image they like best from the first five facial images displayed (e.g., A1, A2, A3, A4, A5).

[1186] Prompt: "Please choose the face image below that you like best."

[1187] Sending user selection information

[1188] For example, if the user selects A3, this information is sent to the server.

[1189] Prompt: "A3 selected. Generate a new face image."

[1190] Generating new face images

[1191] The server generates a set of new face images (e.g., B1, B2, B3, B4, B5) and presents them to the user again.

[1192] Prompt: "Facial images for the next step have been generated. Please choose the one you like best again."

[1193] In this way, the user can intuitively create a facial image of their choice and apply it to a salesperson character in a virtual store.

[1194] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1195] Step 1:

[1196] The server uses an image generation AI model to generate a set of initial facial images and sends their URLs to the device.

[1197] Input: A request comes to the server to generate the initial image.

[1198] Data processing / data calculation: The server generates initial facial images using an image generation AI model, generates a URL for each facial image, and sends it to the device.

[1199] Output: URL list of initial face images.

[1200] Step 2:

[1201] The device downloads the image using the URL of the initial face image received from the server and displays it to the user.

[1202] Input: URL list of initial face images sent from the server.

[1203] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[1204] Output: Multiple initial face images shown to the user.

[1205] Step 3:

[1206] The user selects the facial image that he or she likes best from the presented facial images.

[1207] Input: Multiple displayed face images.

[1208] Data processing / data calculation: Users intuitively select the facial image they like best.

[1209] Output: The index of the selected face image.

[1210] Step 4:

[1211] The terminal transmits the index of the face image selected by the user to the server.

[1212] Input: The index of the face image selected by the user.

[1213] Data processing / data calculation: Converts index information into data to be sent to the server.

[1214] Output: Sends the selection index to the server.

[1215] Step 5:

[1216] The server extracts features from the selected facial image and inputs them into an image generation AI model to learn the user's preferences.

[1217] Input: The index of the selected face image.

[1218] Data processing / data calculation: The server extracts the features of the selected facial image (eye shape, nose height, skin color, etc.) and uses an AI model to learn the user's preferences.

[1219] Output: The training results.

[1220] Step 6:

[1221] The server generates a new set of facial images based on the learning results and sends the URL to the terminal.

[1222] Input: Training results.

[1223] Data processing / data calculation: The server generates new facial images using an AI model and sends their URLs to the device.

[1224] Output: A list of URLs for new face images.

[1225] Step 7:

[1226] The device will use the new face image URL to download the image and display it to the user again.

[1227] Input: A list of URLs for new face images.

[1228] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[1229] Output: Multiple new face images that are displayed to the user.

[1230] Step 8:

[1231] Repeat steps 3 through 7 until the user is satisfied with the generation process.

[1232] Input: A new face image.

[1233] Data processing / data calculation: The user again selects the facial image they like best, and a new facial image is generated based on that.

[1234] Output: The final selected face image.

[1235] Step 9:

[1236] The generated final facial image is applied to a clerk character in a virtual store.

[1237] Input: The final selected face image.

[1238] Data processing / data calculation: The server converts the selected facial image into data to be applied to the virtual store clerk character.

[1239] Output: A custom store clerk character to be used in the virtual store.

[1240] Step 10:

[1241] Users interact with store clerk characters generated within the virtual store.

[1242] Input: A virtual store where the clerk character applies.

[1243] Data processing / data calculation: A store clerk character behaves within the virtual store and interacts with the user.

[1244] Output: A user experience interacting with a custom store clerk character.

[1245] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1246] This invention is a system for efficiently generating facial images that suit a user's preferences, and also combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes that present multiple facial images, learn the user's preferences based on the user's selection and the emotions at that time, and generate new facial images.

[1247] System configuration

[1248] server

[1249] It has an image generation AI model.

[1250] An initial face image and a new face image are generated.

[1251] It receives user selection information and emotion data and learns based on that information.

[1252] Terminal

[1253] The device the user operates (e.g., PC, smartphone).

[1254] The facial image sent from the server is displayed to the user.

[1255] Information about the face image selected by the user and emotion data obtained from the emotion engine are sent to the server.

[1256] user

[1257] Select your preferred facial image and provide feedback to the system.

[1258] Emotion data is collected by an emotion engine.

[1259] Emotion Engine

[1260] It recognizes emotions from the user's facial expressions, voice, heart rate, etc. and collects them as data.

[1261] Processing flow

[1262] Presentation of initial face image

[1263] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[1264] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[1265] User choice and sentiment data collection

[1266] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[1267] At the same time, the emotion engine recognizes the user's emotional data (happiness, surprise, interest, etc.) and transmits this information to the server via the device.

[1268] Learning preferences and generating new face images

[1269] The server receives the user's selection data and emotion data and trains an AI model based on the selected facial image features (e.g., eye shape, nose height, skin color, etc.) and emotion data to learn the user's preferences.

[1270] Based on the learning results, the server generates a new set of facial images that more closely match the user's preferences and sends them to the terminal.

[1271] The terminal again displays the new set of face images to the user, preparing for the next selection.

[1272] Iterative Process

[1273] The user again selects one of the newly presented facial images that they like best. This selection information and emotional data are also recorded on the device and sent to the server. At each step, the server repeatedly performs learning based on the user's selection information and emotional data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1274] Final facial image generation

[1275] After multiple selections and learning processes, the server finally generates a facial image that best reflects the user's preferences, taking into account emotional data, and sends this final facial image to the device.

[1276] Fine-tuning and checking

[1277] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[1278] This system allows users to avoid the hassle of inputting prompts and intuitively generate their preferred facial images with high accuracy. Furthermore, by using an emotion engine, the system can also reflect the user's instantaneous emotions, resulting in even more accurate results.

[1279] The processing flow will be explained below.

[1280] Step 1:

[1281] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[1282] Step 2:

[1283] The device displays the initial set of facial images received from the server to the user, arranging them in a layout that allows the user to easily review each image. At the same time, the emotion engine analyzes the user's facial expressions and prepares to collect initial emotion data.

[1284] Step 3:

[1285] The user selects the facial image they like best from the displayed facial images. Once the user has completed their selection, the device records the selected facial image information and sends it to the server. At the same time, the emotion engine recognizes the user's emotion at the time of selection (e.g., joy, surprise, interest, etc.), and this emotion data is also sent to the server via the device.

[1286] Step 4:

[1287] The server receives the user's selection data and emotion data, extracts features (e.g., eye shape, nose height, skin color, etc.) from the selected facial image, and trains an AI model based on the feature information and emotion data, thereby learning the user's preferences.

[1288] Step 5:

[1289] The server generates a new set of facial images that reflect the learned user's preferred features and emotional data. These facial images are adjusted to more closely match the user's preferences. The generated new set of facial images is then sent to the terminal.

[1290] Step 6:

[1291] The device will then display the newly sent face images to the user again, and the interface will be arranged so that the user can select a face image again, while the emotion engine continues to analyze the user's facial expressions and collect emotional data.

[1292] Step 7:

[1293] The user again selects one of the newly presented facial images that they like best. This selection information and emotion data are also recorded on the device and sent to the server. The server then learns again based on the user's selection information and emotion data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1294] Step 8:

[1295] After a predetermined number of selections and learning processes are completed, the server generates a final facial image that best reflects the user's preferences. This final facial image is generated by combining past selections and emotional data. This final facial image is then sent to the terminal.

[1296] Step 9:

[1297] The device displays the final facial image to the user. The user can review the facial image and make minor adjustments (e.g., changing the size of the eyes or adjusting the hairstyle) as needed. The details of the adjustments are also recorded on the device and, if necessary, sent to the server and reflected. The user can review the results of the adjustments and confirm the final facial image.

[1298] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input, learning preferences step by step, and based on real-time emotional data from the emotion engine, highly accurate results can be achieved.

[1299] Example 2

[1300] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1301] Conventional facial image generation systems require users to input specific prompts, making intuitive operation difficult. Furthermore, they are unable to reflect the user's instantaneous emotions, making it difficult to generate highly accurate facial images. Therefore, there is a demand for a system that can efficiently generate facial images that more accurately reflect the user's preferences.

[1302] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1303] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for recognizing and collecting user emotion data, means for learning the user's preferred features based on the collected selection data and emotion data, and means for generating a new facial image that reflects the learned features. This makes it possible to generate a highly accurate facial image that reflects the user's preferences and takes emotion data into consideration.

[1304] A "user" is a person who operates the system, selects a preferred facial image, and provides emotional data.

[1305] The "server" is a computer system that generates facial images using an image generation AI model and learns based on user selection information and emotional data.

[1306] A "terminal" is a device operated by a user (e.g., a PC or smartphone) that displays a facial image sent from a server and transmits the user's selection information and emotional data to the server.

[1307] An "image generation AI model" is an artificial intelligence model that learns a user's preferred features and generates facial images that reflect those features.

[1308] The "initial face images" are a plurality of face images that the system first presents to the user, and are images that are set to cover a wide range of user preferences.

[1309] "Selection information" is data generated when a user selects a preferred face image from among the presented face images.

[1310] "Emotion data" refers to data of emotions recognized from the user's facial expressions, voice, heart rate, etc.

[1311] An "emotion engine" is a system for recognizing user emotions and collecting emotional data.

[1312] "Learning" is the process by which the AI ​​model analyzes the user's preferred features based on the user's selection information and emotional data, and reflects this in future facial image generation.

[1313] The "new facial image" is a facial image that is closer to the user's preferences and is generated based on the user's selection information and emotion data.

[1314] "Fine tuning" is an operation that allows the user to adjust the details of the generated facial image.

[1315] This invention is a system for efficiently generating facial images that suit a user's preferences, and combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes: presenting multiple facial images, learning preferences based on the user's selection and emotions at the time, and generating new facial images.

[1316] Server Configuration

[1317] The server uses an image generation AI model (e.g., StyleGAN2) to generate facial images that reflect the user's preferences. The server generates an initial facial image and a new facial image based on the results of learning the user's preferences. The server also receives the user's selection information and emotional data and uses that information to learn.

[1318] Device configuration

[1319] The terminal is a device (e.g., PC or smartphone) operated by the user, and is responsible for displaying the facial image sent from the server to the user. It also transmits information about the facial image selected by the user and emotion data obtained from the emotion engine to the server.

[1320] User Roles

[1321] The user selects the facial image they like best from the displayed set of facial images and provides feedback to the system. The emotion engine also collects the user's emotional data.

[1322] Emotion engine configuration

[1323] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc. This data, along with the user's selection, is sent via the device to a server to help improve the accuracy of facial image generation.

[1324] Specific examples

[1325] 1. Generate and present an initial face image:

[1326] The server inputs the prompt "Generate 10 types of facial images" into the image generation AI model, and generates 10 types of initial facial images. These images are sent to the device in JPEG format.

[1327] The terminal displays the transmitted image to the user and provides an interface with selection buttons.

[1328] 2. Collecting user choice and sentiment data:

[1329] The user selects the face image they like best from the ten displayed face images. The selection information is recorded on the device and sent to the server in JSON format.

[1330] The emotion engine uses a camera and microphone to collect the user's facial expressions and tone of voice, as well as heart rate data from a heart rate sensor, which are also sent to the server.

[1331] 3. Generate a new face image:

[1332] The server trains an AI model based on the received selection data and emotion data, inputs the prompt "Please generate a new facial image based on the user's preferences," and generates a new facial image.

[1333] The new facial image is again sent to the terminal in JPEG format and displayed to the user.

[1334] 4. Repeat process:

[1335] The user then selects the face image they like best from the newly presented face images, collects the selection information and emotion data in the same way, and sends them to the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1336] 5. Generate the final face image:

[1337] After multiple selections and learning, the server generates a final facial image that best reflects the user's preferences and sends it to the terminal.

[1338] The user can make fine adjustments to the generated facial image, and the results are also recorded on the device and sent to the server if necessary.

[1339] Hardware and software used

[1340] Server: A high-performance computer system (e.g., a server machine equipped with a GPU)

[1341] Device: PC or smartphone

[1342] Image generation AI model: StyleGAN2

[1343] Prompt Sentence Examples

[1344] "Generate 10 different facial images."

[1345] "Generate new facial images based on user preferences."

[1346] "Generate a facial image based on the user's final preferences."

[1347] This system allows users to intuitively operate the system to accurately generate facial images of their choice. In addition, by using an emotion engine, the system can also reflect the user's momentary emotions, resulting in even more accurate results.

[1348] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1349] Step 1:

[1350] The server generates initial face images using an image generation AI model (e.g., StyleGAN2). The server inputs a prompt statement, "Please generate 10 types of face images," into the generative AI model. 10 types of initial face images (JPEG format) are generated as output.

[1351] Step 2:

[1352] The server sends the generated initial face image to the terminal, and as an output, the face image data is transferred to the terminal.

[1353] Step 3:

[1354] The terminal displays the received initial set of face images to the user, arranges the layout, and provides an interface that adds a selection button to each image. As an output, a set of face images is created that is displayed to the user.

[1355] Step 4:

[1356] The user selects one of the displayed face images that he or she likes the most. The face image that the user likes is selected as an input.

[1357] Step 5:

[1358] The device records the user's selection information and sends it to the server. The selection information is sent to the server in JSON format. As an output, the user's selection information is saved on the server.

[1359] Step 6:

[1360] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc., and sends this information to the server via the device. Emotional data is collected as input, and emotional data is stored on the server as output.

[1361] Step 7:

[1362] The server receives the user's selection data and emotion data and stores them in a database. The selection data and emotion data are sent to the server as input, and these data are recorded in the database as output.

[1363] Step 8:

[1364] The server trains an AI model based on the features (e.g., eye shape, nose height, skin color, etc.) and emotion data of the selected facial images. The feature data and emotion data are used as input, and the AI ​​model is updated as output.

[1365] Step 9:

[1366] The server inputs the prompt "Please generate new facial images based on the user's preferences" into the generative AI model and generates a set of new facial images. The new facial images (JPEG format) are generated as output.

[1367] Step 10:

[1368] The server sends the newly generated face images to the terminal. As an output, the face image data is transferred to the terminal.

[1369] Step 11:

[1370] The terminal displays the new set of face images to the user again and prepares for the next selection. The terminal provides an interface that adjusts the layout and adds a selection button again. As an output, a new set of face images is created and displayed to the user.

[1371] Step 12:

[1372] The user then selects the face image they like best from the newly presented set of face images. The preferred face image is selected as input.

[1373] Step 13:

[1374] The device again records the user's selection information and sends it to the server in JSON format. As an output, the user's selection information is saved on the server.

[1375] Step 14:

[1376] The emotion engine again recognizes the user's emotion data and sends this information to the server via the device. Emotion data is collected as input, and stored as output on the server.

[1377] Step 15:

[1378] The server then trains the AI ​​model again based on the user's selection and emotion data. This process is repeated a set number of times (e.g., 5-10 times), using feature data and emotion data as input and further refining the AI ​​model as output.

[1379] Step 16:

[1380] After multiple selections and learning are completed, the server inputs the prompt sentence "Please generate a facial image based on the user's final preferences" into the generative AI model to generate the final facial image. The final facial image (JPEG format) is generated as the output.

[1381] Step 17:

[1382] The server sends the final generated facial image to the terminal. As an output, the final facial image data is transferred to the terminal.

[1383] Step 18:

[1384] The terminal displays the final face image to the user. As an output, a final face image is created that is displayed to the user.

[1385] Step 19:

[1386] The user reviews this facial image and makes fine adjustments as needed, adding the fine adjustment data as input.

[1387] Step 20:

[1388] The terminal records the results of the fine-tuning and transmits them to the server for reflection as needed. As an output, the final face image is fine-tuned.

[1389] (Application example 2)

[1390] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1391] In conventional systems, simply generating a facial image that reflects a user's preferences takes a lot of time and effort, and they tend to ignore the user's emotions, making it difficult to generate optimal advertisements. Furthermore, advertisements are not sufficiently personalized based on facial image preferences, limiting their effectiveness.

[1392] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for presenting a plurality of facial images to the user, means for the user to select from the displayed facial images, and means for collecting data and emotional data based on the user's selection. This makes it possible to learn the user's preferred features based on the collected data and the user's emotional data, and to generate new facial images and personalized advertisements that reflect the learned features. This makes it possible to efficiently generate and distribute advertisements that reflect the user's preferences and emotions, maximizing advertising effectiveness.

[1393] A "system" is a collection of multiple components and processes that function together.

[1394] A "face image" is image data that digitally represents a person's face.

[1395] "Presenting" is the act of visually displaying information to a user.

[1396] A "means" is a device or method for achieving a specific function.

[1397] "Selection" is the act of choosing one option from multiple options.

[1398] "Data collection" is the process of obtaining information about user choices and responses.

[1399] "Emotional data" is digital information that indicates a user's emotional state.

[1400] "Learning" is the process by which artificial intelligence finds patterns based on data and improves its predictive capabilities.

[1401] "Generation" is the act of creating new digital content.

[1402] "Personalized advertising" refers to advertising content that is individually customized based on a user's preferences and characteristics.

[1403] "Fine-tuning" is the act of making minor modifications to the generated content.

[1404] This invention is a system that generates and delivers customized advertisements based on the user's preferences and emotions. The system collects data on the user's choices and emotions at the time, and learns the user's preferences based on this data. It also generates new facial images and personalized advertisements based on the learned data.

[1405] System configuration

[1406] server

[1407] It has an image generation AI model and generates facial images and advertisements.

[1408] It receives data from the emotion engine and uses it to learn user preferences.

[1409] An initial face image and a new face image and advertisement are generated to be displayed to the user.

[1410] Terminal

[1411] The device the user operates (e.g., PC, smartphone).

[1412] A facial image and advertisements are displayed to the user, and information and emotional data selected by the user are sent to a server.

[1413] user

[1414] Select your preferred facial image and provide feedback to the system.

[1415] Emotion data is collected through an emotion engine.

[1416] Hardware and software used

[1417] Smartphone (iOS or Android device)

[1418] Server (Cloud-based: AWS, Google Cloud Platform, etc.)

[1419] Emotion engine (DeepFace, Emotion API, etc.)

[1420] Image generation AI models (GANs, StyleGAN, etc.)

[1421] Language models (GPT-4, BERT, etc.)

[1422] Data processing and calculation

[1423] The server uses an image generation AI model to generate an initial set of facial images and sends them to the smartphone. The smartphone displays the received facial images to the user, allowing the user to select their preferred image. The selection information and data obtained from the emotion engine are then sent to the server. The server uses this data to train the AI ​​model and learn the user's preferences. The learning results are used to generate new facial images and personalized advertisements, which are then sent back to the smartphone.

[1424] Specific examples

[1425] For example, when a user selects their favorite facial image, the selection information is sent along with emotional data indicating "happiness." The server then uses this data to learn the user's preferences and generate the next facial image and advertisement. By repeating this process, highly accurate personalized advertisements are generated.

[1426] Example prompt sentence:

[1427] "Generate the most preferred face image based on the user's preferences. User preference data: {User preference data} Emotion data: {User emotion data}"

[1428] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1429] Step 1:

[1430] The server uses an image generation AI model to generate a set of initial face images (e.g., 10 different face images). These face images are designed to cover a wide range of user preferences. The generated face images are then sent from the server to the device.

[1431] Input: None (default setting)

[1432] Data processing / computation: Uses image generation AI models to generate facial images based on random or predefined patterns

[1433] Output: Initial face images

[1434] Specific behavior:

[1435] The server runs image generation AI models such as GANs and StyleGAN to generate facial images.

[1436] Step 2:

[1437] The device displays the initial set of facial images sent from the server to the user, and arranges the screen layout to make it easy for the user to select an image.

[1438] Input: Face images sent from the server

[1439] Data processing: Adjusting the layout of face images

[1440] Output: Display face image to user

[1441] Specific behavior:

[1442] Facial images are arranged on the device display and the user interface is designed.

[1443] Step 3:

[1444] The user selects the facial image they like best from the displayed images, and the device collects the user's emotional data (happiness, surprise, interest, etc.) through an emotion engine.

[1445] Input: User selects a face image

[1446] Data processing: Collecting selection information and emotion data

[1447] Output: Information and emotion data of selected face images

[1448] Specific behavior:

[1449] The user taps on the face image of their choice on the touchscreen, and at the same time the emotion engine measures emotional data using the camera and microphone.

[1450] Step 4:

[1451] The device sends user selection information and emotion data to the server, which uses the received data to train an AI model and learn user preferences.

[1452] Input: User selection information and emotion data

[1453] Data processing / computation: Training AI models

[1454] Output: Updated AI model

[1455] Specific behavior:

[1456] The server passes the features and emotional data of the selected facial images as input to the AI ​​model, which then uses machine learning algorithms to learn the user's preferences.

[1457] Step 5:

[1458] The server generates a new set of face images and a personalized advertisement based on the learning results, and transmits the generated face images and advertisement to the terminal again.

[1459] Input: Updated AI model

[1460] Data processing / computation: generating new facial images and advertisements

[1461] Output: New face images and personalized ads

[1462] Specific behavior:

[1463] The server then uses the image generation AI model again based on the learned preferred features to generate new facial images and advertisements.

[1464] Step 6:

[1465] The device displays the newly generated set of facial images and personalized advertisements to the user, and prepares the user to make another selection. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1466] Input: New face image and advertisement sent from the server

[1467] Data processing: Adjusting facial images and ad layout

[1468] Output: Redisplay to user

[1469] Specific behavior:

[1470] New facial images and advertisements are displayed on the device display, and the user interface is redesigned.

[1471] Step 7:

[1472] After the final selection, the server generates a final face image and advertisement that reflects the user's preferences to the maximum extent possible, and transmits them to the terminal.

[1473] Input: Multiple selection information and emotion data

[1474] Data processing / calculation: Generation of final face image and advertisement

[1475] Output: Final face image and advertisement

[1476] Specific behavior:

[1477] The server aggregates the data from all selections and generates the final facial image and personalized advertisements.

[1478] Step 8:

[1479] The device displays the final facial image and advertisement to the user, providing the user with the ability to make fine adjustments as needed.

[1480] Input: Final face image and advertisement

[1481] Data processing: Providing a fine-tuning interface

[1482] Output: Fine-tuned face image and advertisement

[1483] Specific behavior:

[1484] The final facial image and advertisement are displayed on the device display, and a user interface for fine-tuning is provided.

[1485] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1486] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1487] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1488] [Fourth embodiment]

[1489] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1490] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1491] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1492] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1493] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1494] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1495] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1496] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1497] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1498] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1499] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1500] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1501] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1502] The present invention provides a system for efficiently generating facial images that suit a user's preferences. The system includes a series of processes that present multiple facial images, learn the user's preferences based on the user's selection, and generate new facial images.

[1503] System configuration

[1504] server

[1505] It has an image generation AI model.

[1506] An initial face image and a new face image are generated.

[1507] It receives user selection information and learns based on that information.

[1508] Terminal

[1509] The device the user operates (e.g., PC, smartphone).

[1510] The facial image sent from the server is displayed to the user.

[1511] The information of the face image selected by the user is sent to the server.

[1512] user

[1513] Select your preferred facial image and provide feedback to the system.

[1514] Processing flow

[1515] Presentation of initial face image

[1516] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device.

[1517] The device displays these facial images to the user.

[1518] User Selection

[1519] The user selects the facial image that he or she likes best from the presented facial images.

[1520] The terminal transmits this selection information to the server.

[1521] Learning preferences

[1522] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[1523] Generating new face images

[1524] The server generates the next stage of facial images based on the learning results and sends them to the terminal again.

[1525] A new face image will be displayed that reflects the user's selection.

[1526] Iterative Process

[1527] The user again selects the face image they like best from the new set of face images, repeating this process multiple times to provide feedback to the system.

[1528] Through this process, the server gradually learns the user's preferences and generates highly accurate facial images.

[1529] Specific examples

[1530] 1. Presenting the initial face image:

[1531] A group of face images (e.g., A1, A2, A3, A4, A5) generated by the server are displayed on the terminal.

[1532] The user selects A3 and the device sends the information to the server.

[1533] 2. Learning preferences:

[1534] The server extracts features from A3's facial images and learns its preferences.

[1535] 3. Generate a new face image:

[1536] The server generates a new set of facial images (e.g., B1, B2, B3, B4, B5) based on the learning results and sends them to the terminal.

[1537] The user again selects B2, which they like best, and the process is repeated.

[1538] 4. Generate the final face image:

[1539] Finally, after multiple rounds of selection and learning, the server generates a facial image that best reflects the user's preferences and sends it to the device.

[1540] The user can review this facial image and make minor adjustments as needed.

[1541] This system allows users to avoid the hassle of prompt input and use intuitive operations to accurately generate their preferred facial images.

[1542] The processing flow will be explained below.

[1543] Step 1:

[1544] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[1545] Step 2:

[1546] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[1547] Step 3:

[1548] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[1549] Step 4:

[1550] The server receives the user's selection data, extracts features from the selected facial image (e.g., eye shape, nose height, skin color, etc.), and trains an AI model based on the feature information, thereby learning the user's preferences.

[1551] Step 5:

[1552] The server generates a new set of facial images that reflect the learned user's preferred features. These facial images are adjusted to more closely match the user's preferences. The newly generated set of facial images is then sent to the device.

[1553] Step 6:

[1554] The device will then display the newly sent face images to the user again, ready for the user to select the face image they like best.

[1555] Step 7:

[1556] The user again selects one of the newly presented face images that he or she likes best. This selection information is also recorded on the device and sent to the server. As in step 4, the user's selection information is again learned by the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1557] Step 8:

[1558] After multiple selections and learning processes are completed, the server finally generates a facial image that best reflects the user's preferences, and sends this final facial image to the device.

[1559] Step 9:

[1560] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[1561] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input and allowing the system to learn step-by-step based on user selection, highly accurate results can be achieved.

[1562] Example 1

[1563] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1564] Conventional facial image generation systems have struggled to efficiently generate the facial image desired by the user, often forcing the user to perform a huge amount of input work. Furthermore, due to a lack of systems that accurately respond to the user's preferences, the generated facial image often does not meet the user's expectations. The objective of this invention is to solve these problems and provide a system that enables users to intuitively operate the system and generate their preferred facial image with high accuracy.

[1565] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1566] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for transmitting the generated facial image to a user terminal, means for displaying the transmitted facial image to the user terminal, and means for the user to select a preferred facial image from the displayed new facial images. This allows the user to avoid the hassle of prompt input and to generate a preferred facial image with high accuracy through intuitive operation.

[1567] "User" refers to an individual who uses the System to select and generate their preferred facial image.

[1568] "Server" refers to a computer system equipped with an image generation AI model that generates and learns facial images based on user selection information.

[1569] "User terminal" refers to a device operated by a user (e.g., a PC or smartphone) that displays the facial image sent from the server and transmits the user's selection to the server.

[1570] "Image generation AI model" refers to an artificial intelligence model (e.g., StyleGAN) that generates facial images using deep learning technology.

[1571] "Facial Image" means an image of a human face that is generated by the system and presented to the user.

[1572] "Selection information" refers to data (such as the image ID and features) generated when a user selects one face image from multiple images.

[1573] "Features" refer to visual elements contained in a facial image (e.g., eye shape, nose height, skin color, etc.).

[1574] "Learning" refers to the process by which the AI ​​model analyzes and incorporates the user's preferences based on the information in the facial image selected by the user.

[1575] "Display" refers to the act of a user device providing a facial image to the user's visual sense.

[1576] "Generation" refers to the process by which an image-generating AI model generates new facial images.

[1577] "Fine-tuning" refers to the operation in which the user modifies the details of the final generated facial image to make it more closely match their own preferences.

[1578] This invention is a system for efficiently generating facial images that suit a user's preferences. This system consists of several main components, each of which performs a specific function. Specifically, it consists of a server, a user terminal, and a generative AI model.

[1579] System configuration

[1580] server

[1581] The server is equipped with an image generation AI model and has the following functions:

[1582] 1. Generate initial face images: Use an image generation AI model (e.g., StyleGAN) to generate a set of initial face images.

[1583] 2. Learning function: Based on the user's selection information, the system analyzes the characteristics of the facial image and learns the user's preferences.

[1584] 3. Generate new face images: The learned features are reflected to generate the next set of face images.

[1585] 4. Sending face image: Send the generated face image to the user terminal.

[1586] User Device

[1587] The user terminal is a device (e.g., PC, smartphone) that receives face images from the server and is operated by the user. It has the following functions:

[1588] 1. Displaying face image: The face image sent from the server is displayed to the user.

[1589] 2. Sending selected information: Send the information of the face image selected by the user to the server.

[1590] user

[1591] Users interact with the system as follows:

[1592] 1. Select a face image: Select the face image you like best from the displayed face images.

[1593] 2. Providing feedback: Giving feedback to the system through multiple selections.

[1594] Hardware and software used

[1595] Hardware: Servers equipped with high-performance GPUs, devices operated by users (PCs, smartphones)

[1596] Software: Generative AI model (e.g., StyleGAN), user interface (web application or mobile app)

[1597] Specific examples

[1598] 1. The server generates a set of initial face images using StyleGAN. For example, five face images (e.g., A1, A2, A3, A4, A5) are generated.

[1599] 2. The server sends the generated initial face images to the user's device.

[1600] 3. The user selects the facial image (e.g., A3) that they like best from the initial facial images displayed on the device. The device then sends this selection information to the server.

[1601] 4. The server analyzes the features of the selected face image (A3) and learns the user's preferences.

[1602] 5. The server generates a new set of face images (e.g., B1, B2, B3, B4, B5) reflecting the learning results and sends them again to the user's device.

[1603] 6. The user again selects the most preferred face image (e.g., B2) from the new set of face images, and the same process is repeated.

[1604] Prompt Sentence Examples

[1605] Design a system that allows a user to select their first preferred facial image and then learns its features to generate new facial images.

[1606] This system allows users to avoid the hassle of prompt input and to intuitively generate their preferred facial images with high accuracy.

[1607] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1608] Step 1:

[1609] The server generates a set of initial face images using an image generation AI model (e.g., StyleGAN), which uses a random latent space vector as input and generates multiple face images as output. Through this process, the system can obtain an initial set of face images (e.g., A1, A2, A3, A4, A5) to present to the user.

[1610] Step 2:

[1611] The server then sends the generated initial face images to the user's device. The generated face image data is used as input, and is sent to the user's device as output. This process enables the initial face images to be displayed on the device operated by the user.

[1612] Step 3:

[1613] The terminal displays the initial face images received from the server to the user. The received face image data is used as input and displayed on the user's screen as output. This process allows the user to check and select from the initial face images.

[1614] Step 4:

[1615] The user selects the face image they like best from the initial face images displayed. The visually displayed face image is used as input, and the ID and feature information of the selected face image are generated as output. This process determines the user's selection information.

[1616] Step 5:

[1617] The terminal sends the selection information of the face image selected by the user to the server. The user's selection information is used as input and sent to the server as output. This process allows the server to receive the user's preferences.

[1618] Step 6:

[1619] The server analyzes the features of the selected facial image based on the received selection information. The selected facial image data is used as input, and a feature vector is generated as output. This process provides the server with information to learn the user's preferred features.

[1620] Step 7:

[1621] The server inputs the analyzed feature vectors into a generative AI model to learn the user's preferences. The feature vectors are used as input, and updated model parameters are obtained as output. This process allows the generative AI model to reflect the user's preferences.

[1622] Step 8:

[1623] The server then generates a new set of facial images based on the learning results. The updated generative AI model is used as input, and new facial images (e.g., B1, B2, B3, B4, B5) are generated as output. This process provides the next phase of facial images to the user.

[1624] Step 9:

[1625] The server then sends the newly generated face images back to the user's device. The newly generated face image data is used as input, and is sent to the user's device as output. This process allows the user to check the new face images.

[1626] Step 10:

[1627] The device then presents a new set of face images to the user, using the new face image data sent from the server as input and displaying it on the user's screen as output. This process allows the user to select a face image again.

[1628] Step 11:

[1629] The user then selects the face image they like best from a set of new face images. The new face image displayed is used as input, and the information about the reselected face image is generated as output. This process is repeated multiple times until the face image closest to the user's preference is generated.

[1630] (Application example 1)

[1631] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1632] Conventional facial image generation systems have difficulty efficiently generating facial images that match a user's preferences. Furthermore, when applying the generated facial images to a virtual store clerk character, the user's preferences are not fully reflected. Therefore, improving the user experience has become a challenge.

[1633] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1634] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for learning the user's preferred features based on the collected data, means for generating a new facial image that reflects the learned features, means for presenting the generated facial images to the user in a virtual store and for the user to select one, and means for applying the newly generated facial image to a store clerk character in the virtual store. This allows the user to intuitively generate a preferred facial image and have it reflected in the store clerk character in the virtual store.

[1635] A "user" is a person who operates the system, selects his / her preferred facial image, and provides the system with the selected information.

[1636] A "face image" is an image with facial features that is presented for selection by the user.

[1637] The "presentation means" is a mechanism for displaying the facial image generated by the server on the terminal operated by the user.

[1638] The "means for making a selection" is a mechanism that provides an interface for a user to select a preferred facial image from a plurality of displayed facial images.

[1639] "Means for collecting data" refers to a mechanism for transmitting user selection information to a server and storing it.

[1640] The "means of learning features" is a mechanism that uses AI models to analyze and learn users' facial image preferences based on collected user selection data.

[1641] The "means for generating a new facial image" is a mechanism for generating a new facial image that reflects the user's preferences based on the learning results.

[1642] A "virtual store" is a store that exists on the Internet or in a virtual space and allows users to browse and purchase products online.

[1643] A "store clerk character" is a virtual character that appears in the virtual store to introduce and recommend products.

[1644] The present invention provides a system that allows users to efficiently generate facial images of their choice and apply the generated facial images to store clerk characters in a virtual store. This system operates mainly on a server and terminals and includes a process for learning and reflecting user preferences.

[1645] System configuration

[1646] server

[1647] It is equipped with an image generation AI model that generates initial and new facial images, and also receives user selection information and learns based on that information.

[1648] Terminal

[1649] A device operated by the user (e.g., PC, smartphone). It displays the facial images sent from the server to the user and sends the information of the facial image selected by the user to the server.

[1650] user

[1651] Select your preferred facial image and provide feedback to the system.

[1652] Processing flow

[1653] Presentation of initial face image

[1654] The server uses an image generation AI model to generate a set of initial facial images and sends them to the device, where the user is shown these facial images.

[1655] User Selection

[1656] The user selects the facial image that they like best from the presented facial images, and the terminal transmits this selection information to the server.

[1657] Learning preferences

[1658] The server extracts features from the selected facial image (eye shape, nose height, skin color, etc.) and inputs them into an AI model to learn the user's preferences.

[1659] Generating new face images

[1660] The server generates the next set of facial images based on the learning results and sends them back to the device, where new facial images reflecting the user's selection are displayed.

[1661] Applies to virtual stores

[1662] The generated facial image is applied to a store clerk character in a virtual store, and the user can interact with this store clerk character in the virtual store.

[1663] Hardware and software used

[1664] Hardware: Smartphone (iOS or Android), cloud server

[1665] Software: Python, PIL (Python Imaging Library), tkinter (for creating GUI)

[1666] Data processing and calculation

[1667] 1. Presentation of initial face images: Obtain the face image group URL generated by the server, download each image, and process it into display data.

[1668] 2. Sending user selection information: Display a dialog using tkinter, get the index of the image selected by the user, and send it to the server.

[1669] 3. Generate new face images: The server obtains the new face image URL and downloads each image.

[1670] 4. Iteration: The basic process is repeated until a facial image that satisfies the user is generated.

[1671] 5. Application in virtual stores: By applying the generated facial image to a clerk character in a virtual store, users can interact with their favorite clerk character.

[1672] Examples of concrete examples and prompts

[1673] Presentation of initial face image

[1674] The user selects the image they like best from the first five facial images displayed (e.g., A1, A2, A3, A4, A5).

[1675] Prompt: "Please choose the face image below that you like best."

[1676] Sending user selection information

[1677] For example, if the user selects A3, this information is sent to the server.

[1678] Prompt: "A3 selected. Generate a new face image."

[1679] Generating new face images

[1680] The server generates a set of new face images (e.g., B1, B2, B3, B4, B5) and presents them to the user again.

[1681] Prompt: "Facial images for the next step have been generated. Please choose the one you like best again."

[1682] In this way, the user can intuitively create a facial image of their choice and apply it to a salesperson character in a virtual store.

[1683] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1684] Step 1:

[1685] The server uses an image generation AI model to generate a set of initial facial images and sends their URLs to the device.

[1686] Input: A request comes to the server to generate the initial image.

[1687] Data processing / data calculation: The server generates initial facial images using an image generation AI model, generates a URL for each facial image, and sends it to the device.

[1688] Output: URL list of initial face images.

[1689] Step 2:

[1690] The device downloads the image using the URL of the initial face image received from the server and displays it to the user.

[1691] Input: URL list of initial face images sent from the server.

[1692] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[1693] Output: Multiple initial face images shown to the user.

[1694] Step 3:

[1695] The user selects the facial image that he or she likes best from the presented facial images.

[1696] Input: Multiple displayed face images.

[1697] Data processing / data calculation: Users intuitively select the facial image they like best.

[1698] Output: The index of the selected face image.

[1699] Step 4:

[1700] The terminal transmits the index of the face image selected by the user to the server.

[1701] Input: The index of the face image selected by the user.

[1702] Data processing / data calculation: Converts index information into data to be sent to the server.

[1703] Output: Sends the selection index to the server.

[1704] Step 5:

[1705] The server extracts features from the selected facial image and inputs them into an image generation AI model to learn the user's preferences.

[1706] Input: The index of the selected face image.

[1707] Data processing / data calculation: The server extracts the features of the selected facial image (eye shape, nose height, skin color, etc.) and uses an AI model to learn the user's preferences.

[1708] Output: The training results.

[1709] Step 6:

[1710] The server generates a new set of facial images based on the learning results and sends the URL to the terminal.

[1711] Input: Training results.

[1712] Data processing / data calculation: The server generates new facial images using an AI model and sends their URLs to the device.

[1713] Output: A list of URLs for new face images.

[1714] Step 7:

[1715] The device will use the new face image URL to download the image and display it to the user again.

[1716] Input: A list of URLs for new face images.

[1717] Data processing / data calculation: The device downloads the face image from each URL and decodes it to display to the user.

[1718] Output: Multiple new face images that are displayed to the user.

[1719] Step 8:

[1720] Repeat steps 3 through 7 until the user is satisfied with the generation process.

[1721] Input: A new face image.

[1722] Data processing / data calculation: The user again selects the facial image they like best, and a new facial image is generated based on that.

[1723] Output: The final selected face image.

[1724] Step 9:

[1725] The generated final facial image is applied to a clerk character in a virtual store.

[1726] Input: The final selected face image.

[1727] Data processing / data calculation: The server converts the selected facial image into data to be applied to the virtual store clerk character.

[1728] Output: A custom store clerk character to be used in the virtual store.

[1729] Step 10:

[1730] Users interact with store clerk characters generated within the virtual store.

[1731] Input: A virtual store where the clerk character applies.

[1732] Data processing / data calculation: A store clerk character behaves within the virtual store and interacts with the user.

[1733] Output: A user experience interacting with a custom store clerk character.

[1734] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1735] This invention is a system for efficiently generating facial images that suit a user's preferences, and also combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes that present multiple facial images, learn the user's preferences based on the user's selection and the emotions at that time, and generate new facial images.

[1736] System configuration

[1737] server

[1738] It has an image generation AI model.

[1739] An initial face image and a new face image are generated.

[1740] It receives user selection information and emotion data and learns based on that information.

[1741] Terminal

[1742] The device the user operates (e.g., PC, smartphone).

[1743] The facial image sent from the server is displayed to the user.

[1744] Information about the face image selected by the user and emotion data obtained from the emotion engine are sent to the server.

[1745] user

[1746] Select your preferred facial image and provide feedback to the system.

[1747] Emotion data is collected by an emotion engine.

[1748] Emotion Engine

[1749] It recognizes emotions from the user's facial expressions, voice, heart rate, etc. and collects them as data.

[1750] Processing flow

[1751] Presentation of initial face image

[1752] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[1753] The terminal displays the initial set of facial images received from the server to the user, arranges the layout so that the user can easily check each image, and provides an interface for the user to select a facial image.

[1754] User choice and sentiment data collection

[1755] The user selects the facial image that they like best from the displayed facial images. Once the user has completed their selection, the selection information is recorded in the terminal and transmitted to the server.

[1756] At the same time, the emotion engine recognizes the user's emotional data (happiness, surprise, interest, etc.) and transmits this information to the server via the device.

[1757] Learning preferences and generating new face images

[1758] The server receives the user's selection data and emotion data and trains an AI model based on the selected facial image features (e.g., eye shape, nose height, skin color, etc.) and emotion data to learn the user's preferences.

[1759] Based on the learning results, the server generates a new set of facial images that more closely match the user's preferences and sends them to the terminal.

[1760] The terminal again displays the new set of face images to the user, preparing for the next selection.

[1761] Iterative Process

[1762] The user again selects one of the newly presented facial images that they like best. This selection information and emotional data are also recorded on the device and sent to the server. At each step, the server repeatedly performs learning based on the user's selection information and emotional data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1763] Final facial image generation

[1764] After multiple selections and learning processes, the server finally generates a facial image that best reflects the user's preferences, taking into account emotional data, and sends this final facial image to the device.

[1765] Fine-tuning and checking

[1766] The device displays the final facial image to the user, who has the option to review it and make fine adjustments if necessary. The results of the adjustments are also recorded on the device and, if necessary, sent to the server for reflection.

[1767] This system allows users to avoid the hassle of inputting prompts and intuitively generate their preferred facial images with high accuracy. Furthermore, by using an emotion engine, the system can also reflect the user's instantaneous emotions, resulting in even more accurate results.

[1768] The processing flow will be explained below.

[1769] Step 1:

[1770] The server uses an image generation AI model to generate an initial set of facial images (e.g., 10 different facial images). The generated facial image set is configured to cover a wide range of user preferences. These facial images are then sent to the device.

[1771] Step 2:

[1772] The device displays the initial set of facial images received from the server to the user, arranging them in a layout that allows the user to easily review each image. At the same time, the emotion engine analyzes the user's facial expressions and prepares to collect initial emotion data.

[1773] Step 3:

[1774] The user selects the facial image they like best from the displayed facial images. Once the user has completed their selection, the device records the selected facial image information and sends it to the server. At the same time, the emotion engine recognizes the user's emotion at the time of selection (e.g., joy, surprise, interest, etc.), and this emotion data is also sent to the server via the device.

[1775] Step 4:

[1776] The server receives the user's selection data and emotion data, extracts features (e.g., eye shape, nose height, skin color, etc.) from the selected facial image, and trains an AI model based on the feature information and emotion data, thereby learning the user's preferences.

[1777] Step 5:

[1778] The server generates a new set of facial images that reflect the learned user's preferred features and emotional data. These facial images are adjusted to more closely match the user's preferences. The generated new set of facial images is then sent to the terminal.

[1779] Step 6:

[1780] The device will then display the newly sent face images to the user again, and the interface will be arranged so that the user can select a face image again, while the emotion engine continues to analyze the user's facial expressions and collect emotional data.

[1781] Step 7:

[1782] The user again selects one of the newly presented facial images that they like best. This selection information and emotion data are also recorded on the device and sent to the server. The server then learns again based on the user's selection information and emotion data. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1783] Step 8:

[1784] After a predetermined number of selections and learning processes are completed, the server generates a final facial image that best reflects the user's preferences. This final facial image is generated by combining past selections and emotional data. This final facial image is then sent to the terminal.

[1785] Step 9:

[1786] The device displays the final facial image to the user. The user can review the facial image and make minor adjustments (e.g., changing the size of the eyes or adjusting the hairstyle) as needed. The details of the adjustments are also recorded on the device and, if necessary, sent to the server and reflected. The user can review the results of the adjustments and confirm the final facial image.

[1787] This series of processing steps allows users to easily and intuitively generate their preferred facial image. By avoiding prompt input, learning preferences step by step, and based on real-time emotional data from the emotion engine, highly accurate results can be achieved.

[1788] Example 2

[1789] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1790] Conventional facial image generation systems require users to input specific prompts, making intuitive operation difficult. Furthermore, they are unable to reflect the user's instantaneous emotions, making it difficult to generate highly accurate facial images. Therefore, there is a demand for a system that can efficiently generate facial images that more accurately reflect the user's preferences.

[1791] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1792] In this invention, the server includes means for presenting a plurality of facial images to a user, means for the user to select from the displayed facial images, means for collecting data based on the user's selection, means for recognizing and collecting user emotion data, means for learning the user's preferred features based on the collected selection data and emotion data, and means for generating a new facial image that reflects the learned features. This makes it possible to generate a highly accurate facial image that reflects the user's preferences and takes emotion data into consideration.

[1793] A "user" is a person who operates the system, selects a preferred facial image, and provides emotional data.

[1794] The "server" is a computer system that generates facial images using an image generation AI model and learns based on user selection information and emotional data.

[1795] A "terminal" is a device operated by a user (e.g., a PC or smartphone) that displays a facial image sent from a server and transmits the user's selection information and emotional data to the server.

[1796] An "image generation AI model" is an artificial intelligence model that learns a user's preferred features and generates facial images that reflect those features.

[1797] The "initial face images" are a plurality of face images that the system first presents to the user, and are images that are set to cover a wide range of user preferences.

[1798] "Selection information" is data generated when a user selects a preferred face image from among the presented face images.

[1799] "Emotion data" refers to data of emotions recognized from the user's facial expressions, voice, heart rate, etc.

[1800] An "emotion engine" is a system for recognizing user emotions and collecting emotional data.

[1801] "Learning" is the process by which the AI ​​model analyzes the user's preferred features based on the user's selection information and emotional data, and reflects this in future facial image generation.

[1802] The "new facial image" is a facial image that is closer to the user's preferences and is generated based on the user's selection information and emotion data.

[1803] "Fine tuning" is an operation that allows the user to adjust the details of the generated facial image.

[1804] This invention is a system for efficiently generating facial images that suit a user's preferences, and combines it with an emotion engine that recognizes the user's emotions. This system involves a series of processes: presenting multiple facial images, learning preferences based on the user's selection and emotions at the time, and generating new facial images.

[1805] Server Configuration

[1806] The server uses an image generation AI model (e.g., StyleGAN2) to generate facial images that reflect the user's preferences. The server generates an initial facial image and a new facial image based on the results of learning the user's preferences. The server also receives the user's selection information and emotional data and uses that information to learn.

[1807] Device configuration

[1808] The terminal is a device (e.g., PC or smartphone) operated by the user, and is responsible for displaying the facial image sent from the server to the user. It also transmits information about the facial image selected by the user and emotion data obtained from the emotion engine to the server.

[1809] User Roles

[1810] The user selects the facial image they like best from the displayed set of facial images and provides feedback to the system. The emotion engine also collects the user's emotional data.

[1811] Emotion engine configuration

[1812] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc. This data, along with the user's selection, is sent via the device to a server to help improve the accuracy of facial image generation.

[1813] Specific examples

[1814] 1. Generate and present an initial face image:

[1815] The server inputs the prompt "Generate 10 types of facial images" into the image generation AI model, and generates 10 types of initial facial images. These images are sent to the device in JPEG format.

[1816] The terminal displays the transmitted image to the user and provides an interface with selection buttons.

[1817] 2. Collecting user choice and sentiment data:

[1818] The user selects the face image they like best from the ten displayed face images. The selection information is recorded on the device and sent to the server in JSON format.

[1819] The emotion engine uses a camera and microphone to collect the user's facial expressions and tone of voice, as well as heart rate data from a heart rate sensor, which are also sent to the server.

[1820] 3. Generate a new face image:

[1821] The server trains an AI model based on the received selection data and emotion data, inputs the prompt "Please generate a new facial image based on the user's preferences," and generates a new facial image.

[1822] The new facial image is again sent to the terminal in JPEG format and displayed to the user.

[1823] 4. Repeat process:

[1824] The user then selects the face image they like best from the newly presented face images, collects the selection information and emotion data in the same way, and sends them to the server. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1825] 5. Generate the final face image:

[1826] After multiple selections and learning, the server generates a final facial image that best reflects the user's preferences and sends it to the terminal.

[1827] The user can make fine adjustments to the generated facial image, and the results are also recorded on the device and sent to the server if necessary.

[1828] Hardware and software used

[1829] Server: A high-performance computer system (e.g., a server machine equipped with a GPU)

[1830] Device: PC or smartphone

[1831] Image generation AI model: StyleGAN2

[1832] Prompt Sentence Examples

[1833] "Generate 10 different facial images."

[1834] "Generate new facial images based on user preferences."

[1835] "Generate a facial image based on the user's final preferences."

[1836] This system allows users to intuitively operate the system to accurately generate facial images of their choice. In addition, by using an emotion engine, the system can also reflect the user's momentary emotions, resulting in even more accurate results.

[1837] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1838] Step 1:

[1839] The server generates initial face images using an image generation AI model (e.g., StyleGAN2). The server inputs a prompt statement, "Please generate 10 types of face images," into the generative AI model. 10 types of initial face images (JPEG format) are generated as output.

[1840] Step 2:

[1841] The server sends the generated initial face image to the terminal, and as an output, the face image data is transferred to the terminal.

[1842] Step 3:

[1843] The terminal displays the received initial set of face images to the user, arranges the layout, and provides an interface that adds a selection button to each image. As an output, a set of face images is created that is displayed to the user.

[1844] Step 4:

[1845] The user selects one of the displayed face images that he or she likes the most. The face image that the user likes is selected as an input.

[1846] Step 5:

[1847] The device records the user's selection information and sends it to the server. The selection information is sent to the server in JSON format. As an output, the user's selection information is saved on the server.

[1848] Step 6:

[1849] The emotion engine recognizes emotions from the user's facial expressions, voice, heart rate, etc., and sends this information to the server via the device. Emotional data is collected as input, and emotional data is stored on the server as output.

[1850] Step 7:

[1851] The server receives the user's selection data and emotion data and stores them in a database. The selection data and emotion data are sent to the server as input, and these data are recorded in the database as output.

[1852] Step 8:

[1853] The server trains an AI model based on the features (e.g., eye shape, nose height, skin color, etc.) and emotion data of the selected facial images. The feature data and emotion data are used as input, and the AI ​​model is updated as output.

[1854] Step 9:

[1855] The server inputs the prompt "Please generate new facial images based on the user's preferences" into the generative AI model and generates a set of new facial images. The new facial images (JPEG format) are generated as output.

[1856] Step 10:

[1857] The server sends the newly generated face images to the terminal. As an output, the face image data is transferred to the terminal.

[1858] Step 11:

[1859] The terminal displays the new set of face images to the user again and prepares for the next selection. The terminal provides an interface that adjusts the layout and adds a selection button again. As an output, a new set of face images is created and displayed to the user.

[1860] Step 12:

[1861] The user then selects the face image they like best from the newly presented set of face images. The preferred face image is selected as input.

[1862] Step 13:

[1863] The device again records the user's selection information and sends it to the server in JSON format. As an output, the user's selection information is saved on the server.

[1864] Step 14:

[1865] The emotion engine again recognizes the user's emotion data and sends this information to the server via the device. Emotion data is collected as input, and stored as output on the server.

[1866] Step 15:

[1867] The server then trains the AI ​​model again based on the user's selection and emotion data. This process is repeated a set number of times (e.g., 5-10 times), using feature data and emotion data as input and further refining the AI ​​model as output.

[1868] Step 16:

[1869] After multiple selections and learning are completed, the server inputs the prompt sentence "Please generate a facial image based on the user's final preferences" into the generative AI model to generate the final facial image. The final facial image (JPEG format) is generated as the output.

[1870] Step 17:

[1871] The server sends the final generated facial image to the terminal. As an output, the final facial image data is transferred to the terminal.

[1872] Step 18:

[1873] The terminal displays the final face image to the user. As an output, a final face image is created that is displayed to the user.

[1874] Step 19:

[1875] The user reviews this facial image and makes fine adjustments as needed, adding the fine adjustment data as input.

[1876] Step 20:

[1877] The terminal records the results of the fine-tuning and transmits them to the server for reflection as needed. As an output, the final face image is fine-tuned.

[1878] (Application example 2)

[1879] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1880] In conventional systems, simply generating a facial image that reflects a user's preferences takes a lot of time and effort, and they tend to ignore the user's emotions, making it difficult to generate optimal advertisements. Furthermore, advertisements are not sufficiently personalized based on facial image preferences, limiting their effectiveness.

[1881] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for presenting a plurality of facial images to the user, means for the user to select from the displayed facial images, and means for collecting data and emotional data based on the user's selection. This makes it possible to learn the user's preferred features based on the collected data and the user's emotional data, and to generate new facial images and personalized advertisements that reflect the learned features. This makes it possible to efficiently generate and distribute advertisements that reflect the user's preferences and emotions, maximizing advertising effectiveness.

[1882] A "system" is a collection of multiple components and processes that function together.

[1883] A "face image" is image data that digitally represents a person's face.

[1884] "Presenting" is the act of visually displaying information to a user.

[1885] A "means" is a device or method for achieving a specific function.

[1886] "Selection" is the act of choosing one option from multiple options.

[1887] "Data collection" is the process of obtaining information about user choices and responses.

[1888] "Emotional data" is digital information that indicates a user's emotional state.

[1889] "Learning" is the process by which artificial intelligence finds patterns based on data and improves its predictive capabilities.

[1890] "Generation" is the act of creating new digital content.

[1891] "Personalized advertising" refers to advertising content that is individually customized based on a user's preferences and characteristics.

[1892] "Fine-tuning" is the act of making minor modifications to the generated content.

[1893] This invention is a system that generates and delivers customized advertisements based on the user's preferences and emotions. The system collects data on the user's choices and emotions at the time, and learns the user's preferences based on this data. It also generates new facial images and personalized advertisements based on the learned data.

[1894] System configuration

[1895] server

[1896] It has an image generation AI model and generates facial images and advertisements.

[1897] It receives data from the emotion engine and uses it to learn user preferences.

[1898] An initial face image and a new face image and advertisement are generated to be displayed to the user.

[1899] Terminal

[1900] The device the user operates (e.g., PC, smartphone).

[1901] A facial image and advertisements are displayed to the user, and information and emotional data selected by the user are sent to a server.

[1902] user

[1903] Select your preferred facial image and provide feedback to the system.

[1904] Emotion data is collected through an emotion engine.

[1905] Hardware and software used

[1906] Smartphone (iOS or Android device)

[1907] Server (Cloud-based: AWS, Google Cloud Platform, etc.)

[1908] Emotion engine (DeepFace, Emotion API, etc.)

[1909] Image generation AI models (GANs, StyleGAN, etc.)

[1910] Language models (GPT-4, BERT, etc.)

[1911] Data processing and calculation

[1912] The server uses an image generation AI model to generate an initial set of facial images and sends them to the smartphone. The smartphone displays the received facial images to the user, allowing the user to select their preferred image. The selection information and data obtained from the emotion engine are then sent to the server. The server uses this data to train the AI ​​model and learn the user's preferences. The learning results are used to generate new facial images and personalized advertisements, which are then sent back to the smartphone.

[1913] Specific examples

[1914] For example, when a user selects their favorite facial image, the selection information is sent along with emotional data indicating "happiness." The server then uses this data to learn the user's preferences and generate the next facial image and advertisement. By repeating this process, highly accurate personalized advertisements are generated.

[1915] Example prompt sentence:

[1916] "Generate the most preferred face image based on the user's preferences. User preference data: {User preference data} Emotion data: {User emotion data}"

[1917] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1918] Step 1:

[1919] The server uses an image generation AI model to generate a set of initial face images (e.g., 10 different face images). These face images are designed to cover a wide range of user preferences. The generated face images are then sent from the server to the device.

[1920] Input: None (default setting)

[1921] Data processing / computation: Uses image generation AI models to generate facial images based on random or predefined patterns

[1922] Output: Initial face images

[1923] Specific behavior:

[1924] The server runs image generation AI models such as GANs and StyleGAN to generate facial images.

[1925] Step 2:

[1926] The device displays the initial set of facial images sent from the server to the user, and arranges the screen layout to make it easy for the user to select an image.

[1927] Input: Face images sent from the server

[1928] Data processing: Adjusting the layout of face images

[1929] Output: Display face image to user

[1930] Specific behavior:

[1931] Facial images are arranged on the device display and the user interface is designed.

[1932] Step 3:

[1933] The user selects the facial image they like best from the displayed images, and the device collects the user's emotional data (happiness, surprise, interest, etc.) through an emotion engine.

[1934] Input: User selects a face image

[1935] Data processing: Collecting selection information and emotion data

[1936] Output: Information and emotion data of selected face images

[1937] Specific behavior:

[1938] The user taps on the face image of their choice on the touchscreen, and at the same time the emotion engine measures emotional data using the camera and microphone.

[1939] Step 4:

[1940] The device sends user selection information and emotion data to the server, which uses the received data to train an AI model and learn user preferences.

[1941] Input: User selection information and emotion data

[1942] Data processing / computation: Training AI models

[1943] Output: Updated AI model

[1944] Specific behavior:

[1945] The server passes the features and emotional data of the selected facial images as input to the AI ​​model, which then uses machine learning algorithms to learn the user's preferences.

[1946] Step 5:

[1947] The server generates a new set of face images and a personalized advertisement based on the learning results, and transmits the generated face images and advertisement to the terminal again.

[1948] Input: Updated AI model

[1949] Data processing / computation: generating new facial images and advertisements

[1950] Output: New face images and personalized ads

[1951] Specific behavior:

[1952] The server then uses the image generation AI model again based on the learned preferred features to generate new facial images and advertisements.

[1953] Step 6:

[1954] The device displays the newly generated set of facial images and personalized advertisements to the user, and prepares the user to make another selection. This process is repeated a predetermined number of times (e.g., 5 to 10 times).

[1955] Input: New face image and advertisement sent from the server

[1956] Data processing: Adjusting facial images and ad layout

[1957] Output: Redisplay to user

[1958] Specific behavior:

[1959] New facial images and advertisements are displayed on the device display, and the user interface is redesigned.

[1960] Step 7:

[1961] After the final selection, the server generates a final face image and advertisement that reflects the user's preferences to the maximum extent possible, and transmits them to the terminal.

[1962] Input: Multiple selection information and emotion data

[1963] Data processing / calculation: Generation of final face image and advertisement

[1964] Output: Final face image and advertisement

[1965] Specific behavior:

[1966] The server aggregates the data from all selections and generates the final facial image and personalized advertisements.

[1967] Step 8:

[1968] The device displays the final facial image and advertisement to the user, providing the user with the ability to make fine adjustments as needed.

[1969] Input: Final face image and advertisement

[1970] Data processing: Providing a fine-tuning interface

[1971] Output: Fine-tuned face image and advertisement

[1972] Specific behavior:

[1973] The final facial image and advertisement are displayed on the device display, and a user interface for fine-tuning is provided.

[1974] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1975] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1976] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1977] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1978] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1979] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1980] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1981] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1982] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1983] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1984] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1985] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1986] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1987] 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.

[1988] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1989] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1990] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1991] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1992] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1993] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1994] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1995] The following is further disclosed regarding the above embodiment.

[1996] (Claim 1)

[1997] a means for presenting a plurality of facial images to a user;

[1998] a means for a user to select from the displayed facial images;

[1999] means of collecting data based on user preferences;

[2000] A means of learning user preferences based on collected data; and

[2001] A means for generating a new face image that reflects the learned features;

[2002] A system including:

[2003] (Claim 2)

[2004] 10. The system of claim 1, further comprising means for allowing a user to repeatedly make selections to gradually generate a preferred facial image.

[2005] (Claim 3)

[2006] 3. The system according to claim 1, further comprising means for allowing a user to make fine adjustments to the generated facial image.

[2007] "Example 1"

[2008] (Claim 1)

[2009] a means for presenting a plurality of facial images to a user;

[2010] a means for a user to select from the displayed facial images;

[2011] means of collecting data based on user preferences;

[2012] A means of learning user preferences based on collected data; and

[2013] A means for generating a new face image that reflects the learned features;

[2014] A means for transmitting the generated facial image to a user terminal;

[2015] a means for displaying the facial image transmitted to the user terminal to the user;

[2016] a means for selecting a preferred face image from the new face images displayed,

[2017] A system including:

[2018] (Claim 2)

[2019] 10. The system of claim 1, further comprising means for allowing a user to repeatedly make selections to gradually generate a preferred facial image.

[2020] (Claim 3)

[2021] 10. The system of claim 1, further comprising means for a user to make fine adjustments to the generated facial image.

[2022] "Application Example 1"

[2023] (Claim 1)

[2024] a means for presenting a plurality of facial images to a user;

[2025] a means for a user to select from the displayed facial images;

[2026] means of collecting data based on user preferences;

[2027] A means of learning user preferences based on collected data; and

[2028] A means for generating a new face image that reflects the learned features;

[2029] A means for presenting the generated face image to a user in a virtual store and allowing the user to select it;

[2030] A means for applying the newly generated facial image to a salesperson character of a virtual store;

[2031] A system including:

[2032] (Claim 2)

[2033] A method for generating a facial image of a user's preference in stages by repeating selections;

[2034] The system of claim 1, further comprising means for applying the generated facial image to a salesperson character in a virtual store.

[2035] (Claim 3)

[2036] A means for the user to fine-tune the generated facial image;

[2037] The system according to claim 1, further comprising means for displaying the generated facial image in the virtual store in real time and reflecting the fine-tuning results.

[2038] "Example 2: Combining Emotion Engines"

[2039] (Claim 1)

[2040] a means for presenting a plurality of facial images to a user;

[2041] a means for a user to select from the displayed facial images;

[2042] means of collecting data based on user preferences;

[2043] a means for recognizing and collecting user emotional data;

[2044] a means for learning user preference characteristics based on the collected selection data and sentiment data;

[2045] A means for generating a new face image that reflects the learned features;

[2046] A system including:

[2047] (Claim 2)

[2048] The system according to claim 1, wherein the system generates facial images that the user prefers in stages by repeating selections.

[2049] (Claim 3)

[2050] 2. The system of claim 1, wherein the generated facial image is fine-tuned by the user.

[2051] "Application example 2 when combining emotion engines"

[2052] (Claim 1)

[2053] a means for presenting a plurality of facial images to a user;

[2054] a means for a user to select from the displayed facial images;

[2055] means of collecting data based on user preferences;

[2056] A means of learning user preference characteristics based on collected data and user sentiment data;

[2057] A means for generating new facial images and personalized advertisements that reflect the learned features;

[2058] A system including:

[2059] (Claim 2)

[2060] 10. The system of claim 1, further comprising means for allowing a user to iteratively select and progressively generate preferred facial images and advertisements.

[2061] (Claim 3)

[2062] 3. The system of claim 1 or claim 2, further comprising means for a user to make fine adjustments to the generated facial images and advertisements. [Explanation of symbols]

[2063] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. a means for presenting a plurality of facial images to a user; a means for a user to select from the displayed facial images; means of collecting data based on user preferences; A means of learning user preferences based on collected data; and A means for generating a new face image that reflects the learned features; A system including:

2. The system according to claim 1 , further comprising means for allowing the user to repeatedly make selections to gradually generate a preferred facial image.

3. 3. The system according to claim 1, further comprising means for allowing a user to make fine adjustments to the generated facial image.

Images