System

Abstract

A system is provided.SOLUTION: A system comprising: means for receiving and pre-processing an image; means for extracting a feature vector from the pre-processed image; means for using the extracted feature vector to determine a likelihood of generating a AI; means for matching the feature vector to existing images in a database and calculating a similarity; means for calculating a reference ratio; and means for generating and sending a report to a user based on the determination and the reference ratio.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] Conventional methods have struggled to efficiently and accurately determine whether image content was generated by AI. In particular, the process of verifying the origin and reference sources of images obtained online by users is cumbersome and requires manual verification, which takes time and effort. Furthermore, there is a lack of methods to quantitatively measure the reference ratio of an image. This has led to the challenge of making it difficult to determine the authenticity and originality of an image. There is a need for a system that can solve these problems and efficiently and accurately determine the origin and reference ratio of an image. [Means for solving the problem]

[0005] This invention provides a system that includes means for receiving and preprocessing image data, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database to calculate similarity, means for calculating reference ratios, and means for generating a report based on the determination results and reference ratios and sending it to the user. Preprocessing involves resizing and noise removal of images, and edge detection and color histograms are used to extract feature vectors, enabling advanced analysis. This system enables efficient and accurate determination of whether an image is AI-generated, identification of the reference source, and analysis of the reference ratio.

[0006] "Image data" is visual information stored in a computer-readable format.

[0007] "Preprocessing" refers to the adjustment of data at an early stage to improve the accuracy of analysis and analysis results.

[0008] A "feature vector" is a set of data that numerically represents specific patterns or characteristics within an image.

[0009] "AI-Generated Likelihood" refers to the probability that an image was generated using artificial intelligence or machine learning algorithms.

[0010] "Decision" refers to determining whether specific conditions or criteria are met based on extracted features and calculation results.

[0011] A "database" is a system for efficiently storing, managing, and searching large amounts of data.

[0012] "Similarity" is a measure of how similar two pieces of data are.

[0013] The "reference ratio" is a ratio indicating the degree to which a particular image references other images.

[0014] A "report" is a document or data compilation of analysis results and research findings.

[0015] A "user" is a person or entity that operates and uses the system. [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] 1. System Configuration

[0038] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[0039] 2. Program Overview

[0040] The server performs a series of analytical processes on the image data uploaded by the user, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, and analysis result report generation. The detailed operation of each step is explained below.

[0041] 3. User Operation

[0042] Users select the image data they wish to analyze and upload it to the server from their device. User operations are performed via an intuitive and simple interface.

[0043] 4. Server Processing Details

[0044] Pretreatment

[0045] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[0046] Feature Extraction

[0047] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[0048] AI generation judgment

[0049] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[0050] Identifying the source of the misappropriation

[0051] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[0052] Calculating the Reference Ratio

[0053] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0054] 5. Generate and send the report

[0055] The server generates a detailed report based on the results of the judgment and the analysis of the reference ratio. This report includes data on the possibility of AI generation, the identification of the source of misuse, and the reference ratio. The generated report is sent from the server to the user, who can view it on their device.

[0056] Specific examples

[0057] For example, if a user uploads an image called "sunset.jpg," the server performs the following processes in order: Receive "sunset.jpg" and perform appropriate resizing and noise reduction. Next, extract feature vectors using edge detection and color histograms. After that, use a machine learning model to determine whether it is AI-generated and compare it with existing images in the database to identify the source. Finally, calculate the reference ratio between "sunset.jpg" and the most similar image, and send a report summarizing all these results to the user.

[0058] In this way, the present invention allows users to easily and efficiently obtain the origin and reference information of an image, making it easier to verify the authenticity of the image and ensure its originality. This system is highly useful in a variety of application fields.

[0059] The processing flow will be explained below.

[0060] Step 1: Upload user image

[0061] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data to the server.

[0062] Step 2: The server receives the image data

[0063] The server receives the image data sent by the user and checks the format of the image data to see if there are any problems.

[0064] Step 3: The server performs image preprocessing

[0065] The server performs preprocessing on the received image data, including image resizing and noise removal, to prepare the image for further analysis.

[0066] Step 4: The server extracts the feature vectors

[0067] The server extracts feature vectors from the preprocessed images, using techniques such as edge detection and color histograms to extract specific patterns and characteristics of the image as numerical data.

[0068] Step 5: The server determines the possibility of AI generation

[0069] The server inputs the extracted feature vectors into a machine learning model, which evaluates the likelihood of AI generation and generates a probability score. The server then uses this probability score to determine whether the image was AI-generated.

[0070] Step 6: The server identifies potential sources of misappropriation

[0071] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[0072] Step 7: The server calculates the referral rate

[0073] The server calculates the reference ratio using the uploaded image and the candidate image for reuse. The reference ratio is a numerical value that indicates the degree to which a specific image references other images.

[0074] Step 8: Server Generates Report

[0075] The server generates a detailed report based on the AI-generated image judgment results and the analysis of the reference ratio, including the probability that the image was generated by AI, the detected potential sources of misappropriation, and the reference ratio.

[0076] Step 9: Server sends report to user

[0077] The server sends the generated report to the user, who receives the report on their device and checks the analysis results.

[0078] Example 1

[0079] 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."

[0080] Conventional image analysis systems have had difficulty efficiently determining whether image data uploaded by users is AI-generated, identifying the source of misuse, and calculating the reference ratio. In particular, it has been difficult to automatically and accurately perform a series of processes, from highly accurate preprocessing, feature extraction, AI generation determination, reference ratio calculation, and report generation. The purpose of this invention is to address these challenges and provide a system that allows users to easily verify the authenticity and originality of image data.

[0081] 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.

[0082] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This makes it possible to accurately and quickly determine the possibility of AI generation of image data uploaded by a user and identify the source of the misuse.

[0083] "Image data" refers to a digital file format that electronically stores visual information.

[0084] "Preprocessing" refers to initial processing such as resizing and noise removal performed on image data to improve the accuracy of analysis.

[0085] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[0086] "AI generation determination" refers to the process of using a machine learning model to determine whether an image was generated by artificial intelligence.

[0087] "Similarity" is a value calculated by calculating the degree of match between the feature vectors of two images.

[0088] The "reference ratio" is a quantitative ratio that indicates the extent to which an uploaded image references an existing image.

[0089] A "report" is a document summarizing the results of analysis such as the possibility of AI generation, identification of the source of misuse, and reference ratio.

[0090] A "database" is a collection of information that registers and manages a large number of known image data.

[0091] A "machine learning model" is an algorithm that is trained on a large amount of data in advance and then uses the results to analyze new data.

[0092] "Edge detection" is an image processing technique that identifies edges and boundaries within an image.

[0093] A "color histogram" is statistical data that represents the distribution of colors contained in an image.

[0094] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[0095] Hardware and software used

[0096] The following hardware and software are used to implement the present invention.

[0097] Hardware: Server machine equipped with high-speed CPU and GPU

[0098] Software: Image processing libraries (e.g., OpenCV), machine learning libraries (e.g., TensorFlow, PyTorch), database management systems (e.g., MySQL)

[0099] Example of operation

[0100] For example, if a user wants to upload an image called "sunset.jpg", the process is as follows:

[0101] 1. User operations

[0102] The user selects an image file "sunset.jpg" and uploads it to the server via the terminal. An example prompt is as follows:

[0103] Input image: sunset.jpg

[0104] 2. Server Reception

[0105] The server receives the uploaded "sunset.jpg" file and saves it in the specified directory. A message indicating that the file has been received is recorded in the server log.

[0106] 3. Pretreatment

[0107] The server performs preprocessing on the received image files, including image resizing and noise removal. Specifically, the server uses OpenCV to resize the image to a specific size and remove noise.

[0108] 4. Feature Extraction

[0109] After preprocessing, the server extracts feature vectors from the images using edge detection and color histograms, which represent specific patterns and characteristics of the image as numerical data.

[0110] 5. AI generation judgment

[0111] The extracted feature vectors are input into a machine learning model to determine whether the image is AI-generated or not, using a model trained in TensorFlow or PyTorch to obtain a probability score for the likelihood of AI generation.

[0112] 6. Identifying the source of the misappropriation

[0113] The server compares the feature vectors with existing images in a database, which contains a large number of known image data, and identifies the closest candidates based on the calculated similarity.

[0114] 7. Calculation of Reference Ratio

[0115] For each identified source image, the server calculates the reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0116] 8. Generate and send reports

[0117] The server compiles information on the possibility of AI generation, the identification of the source of misuse, and the reference ratio to generate a detailed report. This report is generated in PDF or HTML format and sent to the user, who can then view the analysis results on their device.

[0118] The present invention provides a system that can accurately identify the possibility that image data uploaded by users was generated by AI and the source of the misuse, allowing users to easily verify the authenticity and originality of the image.

[0119] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0120] Step 1: User selects and uploads image data

[0121] Users select the image data they wish to analyze from their own device. For example, they select an image file called "sunset.jpg" and upload it to the server via an intuitive interface. They select the "sunset.jpg" file as input and click the "Upload" button. The image data is sent to the server as output.

[0122] Step 2: The server receives the image data.

[0123] The server receives the "sunset.jpg" file uploaded by the user. As input, the server receives the image data sent from the user's device and saves it in the specified directory. Specifically, a message indicating completion of reception is recorded in the server's log. As output, the image data is saved on the server in a state where it can be preprocessed.

[0124] Step 3: Server preprocessing

[0125] The server performs preprocessing on the received image data. It reads the "sunset.jpg" file as input and performs the following operations: it resizes the image to a specific size using the OpenCV library (cv2.resize(image, (256, 256))) and removes noise (cv2.fastNlMeansDenoisingColored(image, None, 10, 10, 7, 21)). The output is the preprocessed image data.

[0126] Step 4: The server performs feature extraction

[0127] After preprocessing, the server extracts a feature vector from the image data. Using the preprocessed image data as input, it performs the following operations: edge detection (cv2.Canny(image, 100, 200)) and color histogram (cv2.calcHist([image], [0, 1, 2], None, [8, 8, 8], [0, 256, 0, 256, 0, 256])). The output is the feature vector of the image.

[0128] Step 5: The server makes the AI ​​generation decision

[0129] The server inputs the extracted feature vector into a machine learning model to determine whether the image was generated by AI. The feature vector is input into the machine learning model as input, and the following process is performed: Specifically, the feature vector is analyzed using a pre-trained model (e.g., TensorFlow or PyTorch) (model(features)). The output is a probability score for the likelihood of AI generation.

[0130] Step 6: The server identifies the source of the stolen data

[0131] The server compares the extracted feature vector with existing images in the database. As input, it uses the feature vector to calculate the similarity with existing images in the database. Specifically, it performs the similarity calculation and identifies the candidate source with the closest match (compare_with_database(features)). As output, it identifies the most similar existing image.

[0132] Step 7: The server calculates the referral rate.

[0133] Based on the identified source image, the server calculates the reference ratio. Using the feature vectors of the identified source image and the uploaded image data as input, the server executes the following process. Specifically, it calculates the reference ratio with the source image (calculate_reference_ratio(features, best_match_features)). The output shows the reference ratio quantitatively.

[0134] Step 8: The server generates the report and sends it to the user

[0135] The server generates a detailed report based on the analysis results and sends it to the user. It uses the AI-generated probability score, misappropriation source identification, and reference ratio data as input and performs the following process: Specifically, it generates a report in PDF or HTML format (generate_report(probability_score, best_match_details, reference_ratio)) and sends it to the user (send_report_to_user(report)). The analysis results are provided to the user as output.

[0136] (Application example 1)

[0137] 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."

[0138] In modern society, the ease with which counterfeit and inappropriate images are spread over the Internet is increasing, and these false information and inappropriate images can cause serious problems for individuals, companies, and organizations. Therefore, there is a need to quickly and accurately determine the authenticity of images and take appropriate countermeasures as necessary. However, with current technology, determining the authenticity of images and detecting counterfeits often requires time and effort, making real-time detection difficult. This presents a challenge in terms of making it difficult to respond quickly and effectively.

[0139] 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.

[0140] In this invention, the server includes means for receiving and preprocessing image data, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for determining whether the image data is forged, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This enables users to quickly and accurately determine whether images on the Internet are forged and take appropriate measures in real time based on the results.

[0141] "Image data" is visual information represented in digital form.

[0142] "Preprocessing" refers to preparatory work to improve the accuracy of analysis, such as resizing image data and removing noise.

[0143] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[0144] "Likelihood of AI Generation" is a probability score that indicates whether an image was generated by artificial intelligence.

[0145] The "similarity" is a value calculated by using feature vectors to determine the similarity between different images.

[0146] The "reference ratio" is a quantitative index that indicates the degree to which an image references an existing image.

[0147] "Counterfeit determination" is the process of assessing whether an image is counterfeit.

[0148] A "report" is a document summarizing the analysis results and is information provided to the user.

[0149] A "server" is a computer system that processes data on a network.

[0150] A "user" is an individual or organization that uses the system to upload image data and receive analysis results.

[0151] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio. This system is particularly useful for detecting counterfeit images in security services.

[0152] 1. System Configuration

[0153] The present invention mainly comprises the following elements:

[0154] Hardware

[0155] Device: A device that takes and uploads images, such as a smartphone or smart glasses.

[0156] Server: A computer system that receives image data and performs analysis and processing.

[0157] software

[0158] Python: A programming language for writing the overall program logic.

[0159] TensorFlow: A library for extracting feature vectors and running machine learning models to determine AI-generated possibilities.

[0160] Sklearn: A library for computing similarities between feature vectors.

[0161] 2. Specific processing flow

[0162] User operations

[0163] Users use the terminal to select the image data they wish to analyze and upload it to the server. The terminal has an intuitive and simple interface, so users can operate it without any special skills.

[0164] Server Processing

[0165] 1. Pretreatment

[0166] The server receives the image data sent by the user and performs preprocessing such as resizing and noise removal on the received image data.

[0167] 2. Extracting feature vectors

[0168] After preprocessing is complete, feature vectors are extracted from the image data using techniques such as edge detection and color histograms.

[0169] 3. AI-generated judgment

[0170] The extracted feature vector is input into a machine learning model to determine whether the image was generated by AI. The machine learning model is built using TensorFlow and is pre-trained with a large amount of data.

[0171] 4. Identifying the source of the misappropriation

[0172] The feature vector is matched with existing images in a database to identify the most similar images. Similarity is calculated using the cosine_similarity function in Sklearn.

[0173] 5. Calculation of reference ratio

[0174] The reference ratio is calculated based on the similarity between the original image and the uploaded image.

[0175] 6. Report Generation and Notifications

[0176] A detailed report is generated based on the assessment results and reference ratio and sent to the user, who can then check the report on their device and take any necessary measures based on the information.

[0177] Specific examples

[0178] If a user takes an image titled "Important Document.png" during an important company meeting and wants to check whether the image may be forged, they can use this system. The user uploads the image to the server, which performs a series of analysis processes and then provides the user with a determination of whether the image is forged.

[0179] Prompt Sentence Examples

[0180] "Determine if this image was generated by AI, identify which existing images it references, and generate a report on the likelihood that the image is truly original."

[0181] This system allows users to quickly and accurately determine the authenticity of images and take necessary measures in real time.

[0182] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0183] Step 1:

[0184] The user uses the device to select the image data they wish to analyze and upload it to the server. The input is the image data selected by the user operating the device, and the output is the image data sent to the server. Specifically, the user selects "Important Document.png" from the smartphone app and presses the upload button.

[0185] Step 2:

[0186] The server receives the image data sent by the user. The input is the image data sent in step 1, and the output is the image data saved in the server. Specifically, the server receives the HTTP request and saves the image file in a save directory.

[0187] Step 3:

[0188] Preprocessing is performed. The server resizes the received image data and removes noise. The input is the stored image data, and the output is the preprocessed image data. Specifically, the image is resized to 256x256 pixels and noise filtering is applied.

[0189] Step 4:

[0190] Extract feature vectors. The server generates feature vectors from the preprocessed image data using edge detection and color histograms. The input is the preprocessed image data, and the output is a feature vector. Specifically, it extracts edges from the image and quantifies the color distribution using a color histogram.

[0191] Step 5:

[0192] Determine the possibility of AI generation. The server inputs the feature vector into a machine learning model to determine whether the image was generated by AI. The input is the feature vector, and the output is a probability score of AI generation. Specifically, the feature vector is input into the TensorFlow model and the determination result is obtained.

[0193] Step 6:

[0194] The source of the image is identified. The server compares the feature vector with images in an existing database and identifies the most similar image. The input is the feature vector, and the output is the most similar existing image and its similarity. Specifically, the similarity is calculated using Sklearn's cosine_similarity function, and the index of the maximum value is obtained.

[0195] Step 7:

[0196] Calculate the reference ratio. The server calculates the reference ratio based on the similarity with the identified source image. The input is the similarity with the source image, and the output is the reference ratio. Specifically, the server expresses the reference ratio as a percentage based on the similarity score.

[0197] Step 8:

[0198] A report is generated and sent to the user. The server creates a detailed report based on the judgment results and reference ratio and sends it to the user's device. The input is the AI-generated likelihood score and reference ratio, and the output is report data. Specifically, the analysis results are summarized in text format and sent to the user via email or app notification.

[0199] This allows users to immediately check the analysis results and determine the authenticity of the image.

[0200] 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.

[0201] 1. System Configuration

[0202] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.It also combines an emotion engine that recognizes the user's emotions.

[0203] 2. Program Overview

[0204] The server performs a series of analytical processes on image data uploaded by users, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, analysis result report generation, and user emotion recognition using an emotion engine. The detailed operation of each step is explained below.

[0205] 3. User Operation

[0206] The user selects the image data they wish to analyze and uploads it to the server from their device. The device also captures the user's facial expressions and voice data for emotion recognition. User operation is performed via an intuitive and simple interface.

[0207] 4. Server Processing Details

[0208] Pretreatment

[0209] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[0210] Feature Extraction

[0211] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[0212] AI generation judgment

[0213] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[0214] Identifying the source of the misappropriation

[0215] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[0216] Calculating the Reference Ratio

[0217] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0218] 5. Emotion Engine Processing

[0219] Capturing Emotional Data

[0220] The device captures the user's facial and voice data and sends it to the emotion engine, which then analyzes the user's facial and voice data to recognize their emotional state.

[0221] Emotion analysis

[0222] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, and this emotion data is reflected in the analysis results.

[0223] 6. Generate and send reports

[0224] The server generates a detailed report based on the judgment results, reference ratio, and the emotion engine's recognition results. The report includes the possibility that the image was generated by AI, the detected source of misappropriation, the reference ratio, and a corresponding message based on the user's emotion. The report output format is adjusted taking this data into account.

[0225] Specific examples

[0226] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following process: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms. Then, use a machine learning model to determine whether the image is AI-generated and match it with existing images in the database to identify the source. Calculate the reference ratio and generate a detailed report. Finally, the report output format is adjusted based on the user's emotional state of excitement, and sent from the server to the user.

[0227] In this way, by combining the emotion engine, flexible report creation and notification based on the user's emotions becomes possible, making the way analysis results are received more user-friendly. This system is highly useful in a variety of fields.

[0228] The processing flow will be explained below.

[0229] Step 1: Upload user image

[0230] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data along with the user's facial expression and voice data to the server.

[0231] Step 2: The server receives the image data

[0232] The server receives the image data sent by the user, checks the format of the image data, and determines whether it can be analyzed.

[0233] Step 3: The server performs image preprocessing

[0234] The server performs preprocessing on the received image data. This includes image resizing and noise removal. Preprocessing is a necessary preparation to improve the accuracy of analysis.

[0235] Step 4: The server extracts the feature vectors

[0236] The server uses techniques such as edge detection and color histograms to extract feature vectors from the pre-processed images, which represent specific patterns or characteristics of the image as numerical data.

[0237] Step 5: The server determines the possibility of AI generation

[0238] The server uses the extracted feature vector to input data into a machine learning model, which evaluates the likelihood of AI generation and calculates a probability score. Based on this probability score, the server determines whether the image is AI-generated.

[0239] Step 6: The server identifies potential sources of misappropriation

[0240] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[0241] Step 7: The server calculates the referral rate

[0242] The server compares the uploaded image with the potential source image and calculates the reference ratio, which clarifies the extent to which the uploaded image references other images.

[0243] Step 8: The device captures emotion data

[0244] The device captures the user's facial expressions and voice data and sends it to the emotion engine, which allows the user's emotional state to be analyzed in real time.

[0245] Step 9: The server analyzes the emotion data

[0246] The server uses an emotion engine to analyze facial expressions and voice data sent from the device, and classifies the user's emotions into categories such as "happiness," "surprise," "sadness," and "anger."

[0247] Step 10: Server Generates Report

[0248] The server generates a detailed report based on the AI-generated judgment results, the reference ratio, and the emotion engine's recognition results, including the possibility that the image was generated by AI, the detected source of misappropriation, the reference ratio, and a corresponding message according to the user's emotion.

[0249] Step 11: Server sends report to user

[0250] The server sends the generated report to the user, who receives the report on their device and checks the analysis results and corresponding messages.

[0251] Example 2

[0252] 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."

[0253] Conventional image analysis systems can perform preprocessing of image data, extraction of feature vectors, determination of the possibility of AI generation, calculation of similarity, and calculation of reference ratios, but they lack the functionality to reflect the user's emotional state in the analysis results, which means that the user experience is not sufficiently improved.Another problem is that the reports generated based on the analysis results are uniform, making it impossible to respond flexibly to the user's emotions.

[0254] 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.

[0255] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for generating a report based on the determination result and the reference ratio and sending it to the user, means for receiving the user's facial expression and voice data and recognizing their emotional state, and means for adjusting the report output format based on the recognized emotional state. This makes it possible to generate reports that reflect the user's emotional state, making it more user-friendly to receive analysis results.

[0256] "Image data" means visual information represented in digital form.

[0257] "Preprocessing" refers to initial data processing operations performed to improve the accuracy of the analysis of image data.

[0258] A "feature vector" is a numerical representation of characteristics and patterns extracted from image data.

[0259] "AI-generated likelihood determination" is the process of using the extracted feature vectors to determine whether an image was generated by artificial intelligence.

[0260] "Calculation of similarity" is a process of numerically evaluating the similarity between image data and existing image data based on the feature vector.

[0261] The "reference ratio" is a ratio that quantitatively indicates to what extent the uploaded image references existing image data.

[0262] "Report generation" is the process of creating a detailed report based on the analysis results to provide to the user.

[0263] "User's facial expression and voice data" refers to data that digitally represents the user's facial expression and vocal content.

[0264] "Emotional state recognition" is the process of analyzing a user's facial expressions and voice data to determine the user's emotional state.

[0265] "Adjusting report output format" refers to flexibly changing the appearance and content of generated reports based on the perceived emotional state.

[0266] In this system, users upload image data to a server using their devices, and the server analyzes the image data to identify the possibility of AI generation, the reference source, and the reference ratio. Furthermore, it combines an emotion engine that recognizes the user's emotions.

[0267] 1. System Configuration

[0268] The system consists of the following main components:

[0269] Terminal

[0270] server

[0271] Emotion Engine

[0272] 2. User Operation

[0273] Users select the image data they wish to analyze and upload it to the server from their device. In addition to the image data, the device also captures the user's facial expressions and voice data and sends this data to the server. Users operate the system via an intuitive and simple interface.

[0274] 3. Image data processing details (server operation)

[0275] The server processes the received image data in the following steps:

[0276] 1. Pretreatment

[0277] The server resizes the images and removes noise, specifically using libraries such as OpenCV.

[0278] 2. Extracting feature vectors

[0279] Extract feature vectors from preprocessed image data using edge detection (e.g., Canny method) and color histograms. Use functions such as cv2.calcHist.

[0280] 3. AI-generated possibility assessment

[0281] The feature vector is fed into a machine learning model (e.g., a TensorFlow model) to assess the likelihood that the image was generated by AI.

[0282] 4. Identifying the source of the misappropriation

[0283] The feature vectors are matched with existing images in a database, and potential sources are identified based on similarity (e.g., cosine similarity).

[0284] 5. Calculation of Reference Ratio

[0285] The reference ratio is calculated for the original image, and the extent to which the existing image is referenced is quantitatively shown.

[0286] 4. Emotion engine processing (terminal and server operation)

[0287] 1. Capturing Emotional Data

[0288] The device captures the user's facial expressions and voice data and sends that data to the emotion engine.

[0289] 2. Emotion analysis

[0290] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, using facial recognition algorithms and voice emotion recognition models.

[0291] 5. Generate and send reports (server operations)

[0292] The server generates a detailed report based on the judgment results, reference ratios, and the emotion engine's recognition results, including:

[0293] The image may have been generated by AI

[0294] Identified potential sources of misappropriation

[0295] Reference ratio

[0296] Response messages based on user sentiment

[0297] This report is sent to the user, and the output format is adjusted according to the user's feelings.

[0298] Specific examples

[0299] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following processes: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms, and use a machine learning model to determine the possibility of AI generation. Match the image with existing images in the database to identify the source and calculate the reference ratio. Generate a detailed report, adjust the report output format based on the user's emotional state of excitement, and send it to the user from the server.

[0300] Example prompts for generative AI models

[0301] "Please determine whether the uploaded image data was generated by AI and analyze whether it has a specific source. Also, please determine the user's emotion and generate a report based on that. For example, the image file name is 'sunset.jpg' and the user's emotion is 'excited'."

[0302] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0303] Step 1: Upload image data

[0304] Input: An image file selected by the user on the device.

[0305] Processing: The user selects the image data they wish to analyze and uploads it from their device to the server.

[0306] Output: Image file uploaded to the server

[0307] Specific operation: The user selects the desired image file (e.g., "sunset.jpg") from the device's file system and clicks the "Upload" button, which sends the image data to the server.

[0308] Step 2: Capturing emotion data

[0309] Input: User facial and voice data

[0310] Processing: The device captures the user's facial expressions and voice and sends them to the server as emotional data.

[0311] Output: Facial expression and voice data sent to the server

[0312] Specific operation: The device's camera and microphone are used to record and photograph the user's face and voice in real time, and the recorded and photographed data is sent to a server via the network.

[0313] Step 3: Preprocessing the image data

[0314] Input: Image file uploaded to the server

[0315] Processing: The server performs preprocessing on the received image data.

[0316] Output: Preprocessed image data

[0317] Specific operation: The server uses a library (e.g. OpenCV) to resize the image data to a standard size and remove noise using filtering techniques.

[0318] Step 4: Extract feature vectors

[0319] Input: Preprocessed image data

[0320] Processing: The server extracts feature vectors from the preprocessed image data.

[0321] Output: Extracted feature vector

[0322] Specific operation: The server uses edge detection (e.g., Canny method) and color histogram (e.g., cv2.calcHist function) to extract image features as numerical data.

[0323] Step 5: Determine the possibility of AI generation

[0324] Input: Extracted feature vector

[0325] Processing: The server inputs the extracted feature vectors into a machine learning model to determine the likelihood that the image was generated by AI.

[0326] Output: Probability score indicating the likelihood of the AI ​​generation

[0327] What it does: The server uses a trained machine learning model (e.g., a TensorFlow model) to evaluate feature vectors and output an AI-generated probability score.

[0328] Step 6: Identify the source

[0329] Input: Extracted feature vector

[0330] Processing: The server matches the feature vector with existing images in a database to identify potential sources.

[0331] Output: Identified potential sources of misappropriation

[0332] Specific operation: The server queries the database for the feature vector and identifies the closest match based on a calculated similarity (e.g., cosine similarity).

[0333] Step 7: Calculate the reference ratio

[0334] Input: Uploaded image file, identified potential source

[0335] Processing: The server calculates the reference ratio for the identified source image.

[0336] Output: Reference ratio

[0337] Specific operation: The server calculates the reference ratio based on the similarity between the uploaded image and the original image, and quantitatively indicates the extent to which the image has been referenced.

[0338] Step 8: Analyze the sentiment data

[0339] Input: Facial and voice data sent to the server

[0340] Processing: The server's emotion engine analyzes the received facial and voice data to recognize the user's emotional state.

[0341] Output: User's emotional state

[0342] Specific operation: The server uses an emotion analysis algorithm (e.g., a facial recognition algorithm or a voice emotion recognition model) to analyze emotions from the user's facial expressions and voice.

[0343] Step 9: Generate reports

[0344] Input: Probability score indicating the likelihood of AI generation, identified potential sources of misappropriation, reference ratio, and user emotional state

[0345] Processing: The server generates a report based on the judgment results, reference ratio, and the emotion engine's recognition results.

[0346] Output: Detailed report for the user

[0347] Specific operation: The server organizes the analysis results and outputs a report in a format that is easy for the user to understand (e.g., PDF or HTML format). The report includes the possibility of the image being generated by AI, potential sources of reuse, the reference ratio, and a corresponding message based on the user's emotions.

[0348] Step 10: Submit the report

[0349] Input: Detailed reports for users

[0350] Processing: The server generates the report and sends it to the user.

[0351] Output: Report sent to user

[0352] What happens: The server sends the report to the user via email or a download link. The user can download the report by clicking the provided link.

[0353] (Application example 2)

[0354] 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."

[0355] There is a need to verify the originality of image materials used by advertising creators and optimize ad placement based on user emotions, but conventional systems have had difficulty doing this effectively.In addition, there was no system that could flexibly respond by taking user emotions into account when evaluating the possibility that an image was generated by AI or identifying the source of the image.

[0356] 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 receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for capturing user emotion data and adjusting the analysis results based on the emotion data, and means for generating a report based on the determination results, reference ratio, and emotion data and sending it to the user. This makes it possible to confirm the originality of image materials used by advertising creators, evaluate the possibility of AI generation, and identify sources of reuse, as well as to suggest optimal ad placement based on user emotion.

[0357] "Image data" is digital data of visual information that a user uploads to a server using a terminal.

[0358] "Preprocessing" refers to initial processing such as resizing and noise removal that is performed before analyzing image data.

[0359] A "feature vector" is a numerical representation of a particular pattern or characteristic of image data.

[0360] "Likelihood of AI Generation" is a probability score that assesses whether image data was generated by artificial intelligence.

[0361] "Similarity" is a numerical representation of the degree of match between feature vectors.

[0362] The "reference ratio" is the percentage that indicates the extent to which the uploaded image references an existing image.

[0363] "Emotion data" is information that indicates the emotional state of the user analyzed from facial expressions and voice data.

[0364] "Capture" refers to the device recording the user's facial expressions and voice and acquiring them as digital data.

[0365] The "analysis results" are a compilation of all the processing results from the server's preprocessing of image data, feature extraction, AI generation judgment, similarity calculation, reference ratio calculation, and emotion analysis.

[0366] A "report" is a detailed report generated based on the analysis results and emotion data and provided to the user.

[0367] 1. System Configuration

[0368] This invention is a system aimed at verifying the originality of image materials used by advertising creators and optimizing advertising placement based on user emotions. The system consists of three main elements: a server, a terminal, and a user. The server analyzes image data, the terminal captures image data and emotional data from the user, and the user uploads the image data.

[0369] 2. Program Overview

[0370] To implement this invention, the following hardware and software are used. The hardware requires a terminal equipped with a camera for capturing images and a microphone for capturing audio, and a server capable of high-speed computation. The software requires a library using OpenCV (cv2) to resize and denoise images, an engine using EmotionEngine to recognize user emotions, a module using FeatureExtractor to extract feature vectors, a module using AIDetector to determine the possibility of AI generation, a module using SimilaritySearch to identify the source of reuse, and a module using ReportGenerator to generate the final report.

[0371] 3. Processing Flow

[0372] The user selects the image data they wish to analyze and uploads it from their device to the server. The device simultaneously captures the user's facial expressions and voice data and transmits their emotional state to the server. The server first performs preprocessing on the received image data, including image resizing and noise removal. After the preprocessing is complete, feature vectors are extracted from the image data. Edge detection and color histograms are used to extract the feature vectors.

[0373] The server then uses the extracted feature vectors to assess the likelihood of AI generation. This is done using a machine learning model previously trained on a large amount of data. The result of the AI ​​generation assessment is a probability score indicating whether the image was generated by AI. The server then matches the feature vectors with existing images in its database to identify the source of the image. For each identified source image, it calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references existing images.

[0374] Furthermore, the emotion engine analyzes the user's emotional state and reflects this emotional data in the analysis results. Finally, the server generates a detailed report based on the analysis results, reference ratio, and emotion data and sends it to the user. The report includes the possibility that the image was generated by AI, the original image and reference ratio, and a corresponding message based on the user's emotion.

[0375] 4. Specific Examples

[0376] For example, consider the case where an ad creator uploads a new ad banner image "ad_banner.jpg" to the app and simultaneously captures user sentiment data (interestingness). The server receives this image, resizes it, and denoises it. Next, edge detection and color histograms are used to extract feature vectors. A machine learning model is then used to determine whether it is AI-generated and match it with existing images in the database to identify the source. The reference ratio is calculated and a detailed report is generated. Finally, a report containing optimal ad placement suggestions based on user sentiment is generated and sent to the creator.

[0377] Example prompt sentence:

[0378] "Upload your ad banner image and see how original and AI-generated it is."

[0379] "Please suggest the best ad placement if the user finds it interesting."

[0380] In this way, this system can verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose optimal ad placements based on user sentiment.

[0381] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0382] Step 1:

[0383] The user selects the image data they wish to analyze and uploads it from their device to the server. The input is the image data "ad_banner.jpg" selected by the user, and the output is the image data sent to the server. In this process, the device reads the image data and sends it to the specified server's upload endpoint.

[0384] Step 2:

[0385] The device captures the user's facial and voice data and sends the emotion data to the server. The input is the user's facial and voice data, and the output is the emotion data sent to the server. In this process, the device captures data in real time using a camera and microphone and sends it to the server in a format for emotion analysis.

[0386] Step 3:

[0387] The server performs preprocessing on the received image data. The input is the image data "ad_banner.jpg" sent to the server, and the output is the image data after resizing and noise removal. In this process, OpenCV (cv2) is used to resize the image to 256x256 pixels and apply a noise removal algorithm.

[0388] Step 4:

[0389] The server extracts feature vectors from the preprocessed image data. The input is the preprocessed image data, and the output is a feature vector. In this process, FeatureExtractor is used to apply edge detection and color histogram techniques to extract image characteristics as numerical data.

[0390] Step 5:

[0391] The server uses the extracted feature vector to determine the possibility of AI generation. The input is the extracted feature vector, and the output is a probability score indicating the possibility of AI generation. In this process, the feature vector is input into the pre-trained AIDetector machine learning model to evaluate the possibility of AI generation.

[0392] Step 6:

[0393] The server compares the feature vector with existing images in the database and calculates the similarity. The input is the extracted feature vector, and the output is the similarity score with the existing image and the candidate image to be reused. In this process, SimilaritySearch is used to compare the feature vector with images in the database and identify the most similar image.

[0394] Step 7:

[0395] The server calculates the reference ratio. The input is the similarity score with existing images, and the output is the reference ratio. This process quantitatively calculates how much the uploaded image references existing images based on the similarity score.

[0396] Step 8:

[0397] The server analyzes the user's emotional data and adjusts the analysis results based on that emotional state. The input is the emotional data, and the output is the adjusted analysis results. In this process, the Emotion Engine is used to recognize the user's emotional state, and the analysis results reflect responses that match those emotions.

[0398] Step 9:

[0399] The server generates a report based on the judgment results, reference ratios, and emotion data and sends it to the user. The inputs are the judgment results, reference ratios, and emotion data, and the output is the generated report. In this process, the ReportGenerator aggregates the analysis results and creates and notifies the user of the report in an easy-to-understand format.

[0400] Through each of the above steps, this system is able to verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose appropriate ad placements based on user sentiment.

[0401] 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.

[0402] 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.

[0403] 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.

[0404] [Second embodiment]

[0405] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0406] 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.

[0407] 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).

[0408] 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.

[0409] 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.

[0410] 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).

[0411] 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.

[0412] 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.

[0413] 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.

[0414] 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.

[0415] 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.

[0416] 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."

[0417] 1. System Configuration

[0418] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[0419] 2. Program Overview

[0420] The server performs a series of analytical processes on the image data uploaded by the user, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, and analysis result report generation. The detailed operation of each step is explained below.

[0421] 3. User Operation

[0422] Users select the image data they wish to analyze and upload it to the server from their device. User operations are performed via an intuitive and simple interface.

[0423] 4. Server Processing Details

[0424] Pretreatment

[0425] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[0426] Feature Extraction

[0427] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[0428] AI generation judgment

[0429] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[0430] Identifying the source of the misappropriation

[0431] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[0432] Calculating the Reference Ratio

[0433] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0434] 5. Generate and send the report

[0435] The server generates a detailed report based on the results of the judgment and the analysis of the reference ratio. This report includes data on the possibility of AI generation, the identification of the source of misuse, and the reference ratio. The generated report is sent from the server to the user, who can view it on their device.

[0436] Specific examples

[0437] For example, if a user uploads an image called "sunset.jpg," the server performs the following processes in order: Receive "sunset.jpg" and perform appropriate resizing and noise reduction. Next, extract feature vectors using edge detection and color histograms. After that, use a machine learning model to determine whether it is AI-generated and compare it with existing images in the database to identify the source. Finally, calculate the reference ratio between "sunset.jpg" and the most similar image, and send a report summarizing all these results to the user.

[0438] In this way, the present invention allows users to easily and efficiently obtain the origin and reference information of an image, making it easier to verify the authenticity of the image and ensure its originality. This system is highly useful in a variety of application fields.

[0439] The processing flow will be explained below.

[0440] Step 1: Upload user image

[0441] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data to the server.

[0442] Step 2: The server receives the image data

[0443] The server receives the image data sent by the user and checks the format of the image data to see if there are any problems.

[0444] Step 3: The server performs image preprocessing

[0445] The server performs preprocessing on the received image data, including image resizing and noise removal, to prepare the image for further analysis.

[0446] Step 4: The server extracts the feature vectors

[0447] The server extracts feature vectors from the preprocessed images, using techniques such as edge detection and color histograms to extract specific patterns and characteristics of the image as numerical data.

[0448] Step 5: The server determines the possibility of AI generation

[0449] The server inputs the extracted feature vectors into a machine learning model, which evaluates the likelihood of AI generation and generates a probability score. The server then uses this probability score to determine whether the image was AI-generated.

[0450] Step 6: The server identifies potential sources of misappropriation

[0451] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[0452] Step 7: The server calculates the referral rate

[0453] The server calculates the reference ratio using the uploaded image and the candidate image for reuse. The reference ratio is a numerical value that indicates the degree to which a specific image references other images.

[0454] Step 8: Server Generates Report

[0455] The server generates a detailed report based on the AI-generated image judgment results and the analysis of the reference ratio, including the probability that the image was generated by AI, the detected potential sources of misappropriation, and the reference ratio.

[0456] Step 9: Server sends report to user

[0457] The server sends the generated report to the user, who receives the report on their device and checks the analysis results.

[0458] Example 1

[0459] 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."

[0460] Conventional image analysis systems have had difficulty efficiently determining whether image data uploaded by users is AI-generated, identifying the source of misuse, and calculating the reference ratio. In particular, it has been difficult to automatically and accurately perform a series of processes, from highly accurate preprocessing, feature extraction, AI generation determination, reference ratio calculation, and report generation. The purpose of this invention is to address these challenges and provide a system that allows users to easily verify the authenticity and originality of image data.

[0461] 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.

[0462] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This makes it possible to accurately and quickly determine the possibility of AI generation of image data uploaded by a user and identify the source of the misuse.

[0463] "Image data" refers to a digital file format that electronically stores visual information.

[0464] "Preprocessing" refers to initial processing such as resizing and noise removal performed on image data to improve the accuracy of analysis.

[0465] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[0466] "AI generation determination" refers to the process of using a machine learning model to determine whether an image was generated by artificial intelligence.

[0467] "Similarity" is a value calculated by calculating the degree of match between the feature vectors of two images.

[0468] The "reference ratio" is a quantitative ratio that indicates the extent to which an uploaded image references an existing image.

[0469] A "report" is a document summarizing the results of analysis such as the possibility of AI generation, identification of the source of misuse, and reference ratio.

[0470] A "database" is a collection of information that registers and manages a large number of known image data.

[0471] A "machine learning model" is an algorithm that is trained on a large amount of data in advance and then uses the results to analyze new data.

[0472] "Edge detection" is an image processing technique that identifies edges and boundaries within an image.

[0473] A "color histogram" is statistical data that represents the distribution of colors contained in an image.

[0474] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[0475] Hardware and software used

[0476] The following hardware and software are used to implement the present invention.

[0477] Hardware: Server machine equipped with high-speed CPU and GPU

[0478] Software: Image processing libraries (e.g., OpenCV), machine learning libraries (e.g., TensorFlow, PyTorch), database management systems (e.g., MySQL)

[0479] Example of operation

[0480] For example, if a user wants to upload an image called "sunset.jpg", the process is as follows:

[0481] 1. User operations

[0482] The user selects an image file "sunset.jpg" and uploads it to the server via the terminal. An example prompt is as follows:

[0483] Input image: sunset.jpg

[0484] 2. Server Reception

[0485] The server receives the uploaded "sunset.jpg" file and saves it in the specified directory. A message indicating that the file has been received is recorded in the server log.

[0486] 3. Pretreatment

[0487] The server performs preprocessing on the received image files, including image resizing and noise removal. Specifically, the server uses OpenCV to resize the image to a specific size and remove noise.

[0488] 4. Feature Extraction

[0489] After preprocessing, the server extracts feature vectors from the images using edge detection and color histograms, which represent specific patterns and characteristics of the image as numerical data.

[0490] 5. AI generation judgment

[0491] The extracted feature vectors are input into a machine learning model to determine whether the image is AI-generated or not, using a model trained in TensorFlow or PyTorch to obtain a probability score for the likelihood of AI generation.

[0492] 6. Identifying the source of the misappropriation

[0493] The server compares the feature vectors with existing images in a database, which contains a large number of known image data, and identifies the closest candidates based on the calculated similarity.

[0494] 7. Calculation of Reference Ratio

[0495] For each identified source image, the server calculates the reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0496] 8. Generate and send reports

[0497] The server compiles information on the possibility of AI generation, the identification of the source of misuse, and the reference ratio to generate a detailed report. This report is generated in PDF or HTML format and sent to the user, who can then view the analysis results on their device.

[0498] The present invention provides a system that can accurately identify the possibility that image data uploaded by users was generated by AI and the source of the misuse, allowing users to easily verify the authenticity and originality of the image.

[0499] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0500] Step 1: User selects and uploads image data

[0501] Users select the image data they wish to analyze from their own device. For example, they select an image file called "sunset.jpg" and upload it to the server via an intuitive interface. They select the "sunset.jpg" file as input and click the "Upload" button. The image data is sent to the server as output.

[0502] Step 2: The server receives the image data.

[0503] The server receives the "sunset.jpg" file uploaded by the user. As input, the server receives the image data sent from the user's device and saves it in the specified directory. Specifically, a message indicating completion of reception is recorded in the server's log. As output, the image data is saved on the server in a state where it can be preprocessed.

[0504] Step 3: Server preprocessing

[0505] The server performs preprocessing on the received image data. It reads the "sunset.jpg" file as input and performs the following operations: it resizes the image to a specific size using the OpenCV library (cv2.resize(image, (256, 256))) and removes noise (cv2.fastNlMeansDenoisingColored(image, None, 10, 10, 7, 21)). The output is the preprocessed image data.

[0506] Step 4: The server performs feature extraction

[0507] After preprocessing, the server extracts a feature vector from the image data. Using the preprocessed image data as input, it performs the following operations: edge detection (cv2.Canny(image, 100, 200)) and color histogram (cv2.calcHist([image], [0, 1, 2], None, [8, 8, 8], [0, 256, 0, 256, 0, 256])). The output is the feature vector of the image.

[0508] Step 5: The server makes the AI ​​generation decision

[0509] The server inputs the extracted feature vector into a machine learning model to determine whether the image was generated by AI. The feature vector is input into the machine learning model as input, and the following process is performed: Specifically, the feature vector is analyzed using a pre-trained model (e.g., TensorFlow or PyTorch) (model(features)). The output is a probability score for the likelihood of AI generation.

[0510] Step 6: The server identifies the source of the stolen data

[0511] The server compares the extracted feature vector with existing images in the database. As input, it uses the feature vector to calculate the similarity with existing images in the database. Specifically, it performs the similarity calculation and identifies the candidate source with the closest match (compare_with_database(features)). As output, it identifies the most similar existing image.

[0512] Step 7: The server calculates the referral rate.

[0513] Based on the identified source image, the server calculates the reference ratio. Using the feature vectors of the identified source image and the uploaded image data as input, the server executes the following process. Specifically, it calculates the reference ratio with the source image (calculate_reference_ratio(features, best_match_features)). The output shows the reference ratio quantitatively.

[0514] Step 8: The server generates the report and sends it to the user

[0515] The server generates a detailed report based on the analysis results and sends it to the user. It uses the AI-generated probability score, misappropriation source identification, and reference ratio data as input and performs the following process: Specifically, it generates a report in PDF or HTML format (generate_report(probability_score, best_match_details, reference_ratio)) and sends it to the user (send_report_to_user(report)). The analysis results are provided to the user as output.

[0516] (Application example 1)

[0517] 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."

[0518] In modern society, the ease with which counterfeit and inappropriate images are spread over the Internet is increasing, and these false information and inappropriate images can cause serious problems for individuals, companies, and organizations. Therefore, there is a need to quickly and accurately determine the authenticity of images and take appropriate countermeasures as necessary. However, with current technology, determining the authenticity of images and detecting counterfeits often requires time and effort, making real-time detection difficult. This presents a challenge in terms of making it difficult to respond quickly and effectively.

[0519] 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.

[0520] In this invention, the server includes means for receiving and preprocessing image data, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for determining whether the image data is forged, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This enables users to quickly and accurately determine whether images on the Internet are forged and take appropriate measures in real time based on the results.

[0521] "Image data" is visual information represented in digital form.

[0522] "Preprocessing" refers to preparatory work to improve the accuracy of analysis, such as resizing image data and removing noise.

[0523] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[0524] "Likelihood of AI Generation" is a probability score that indicates whether an image was generated by artificial intelligence.

[0525] The "similarity" is a value calculated by using feature vectors to determine the similarity between different images.

[0526] The "reference ratio" is a quantitative index that indicates the degree to which an image references an existing image.

[0527] "Counterfeit determination" is the process of assessing whether an image is counterfeit.

[0528] A "report" is a document summarizing the analysis results and is information provided to the user.

[0529] A "server" is a computer system that processes data on a network.

[0530] A "user" is an individual or organization that uses the system to upload image data and receive analysis results.

[0531] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio. This system is particularly useful for detecting counterfeit images in security services.

[0532] 1. System Configuration

[0533] The present invention mainly comprises the following elements:

[0534] Hardware

[0535] Device: A device that takes and uploads images, such as a smartphone or smart glasses.

[0536] Server: A computer system that receives image data and performs analysis and processing.

[0537] software

[0538] Python: A programming language for writing the overall program logic.

[0539] TensorFlow: A library for extracting feature vectors and running machine learning models to determine AI-generated possibilities.

[0540] Sklearn: A library for computing similarities between feature vectors.

[0541] 2. Specific processing flow

[0542] User operations

[0543] Users use the terminal to select the image data they wish to analyze and upload it to the server. The terminal has an intuitive and simple interface, so users can operate it without any special skills.

[0544] Server Processing

[0545] 1. Pretreatment

[0546] The server receives the image data sent by the user and performs preprocessing such as resizing and noise removal on the received image data.

[0547] 2. Extracting feature vectors

[0548] After preprocessing is complete, feature vectors are extracted from the image data using techniques such as edge detection and color histograms.

[0549] 3. AI-generated judgment

[0550] The extracted feature vector is input into a machine learning model to determine whether the image was generated by AI. The machine learning model is built using TensorFlow and is pre-trained with a large amount of data.

[0551] 4. Identifying the source of the misappropriation

[0552] The feature vector is matched with existing images in a database to identify the most similar images. Similarity is calculated using the cosine_similarity function in Sklearn.

[0553] 5. Calculation of reference ratio

[0554] The reference ratio is calculated based on the similarity between the original image and the uploaded image.

[0555] 6. Report Generation and Notifications

[0556] A detailed report is generated based on the assessment results and reference ratio and sent to the user, who can then check the report on their device and take any necessary measures based on the information.

[0557] Specific examples

[0558] If a user takes an image titled "Important Document.png" during an important company meeting and wants to check whether the image may be forged, they can use this system. The user uploads the image to the server, which performs a series of analysis processes and then provides the user with a determination of whether the image is forged.

[0559] Prompt Sentence Examples

[0560] "Determine if this image was generated by AI, identify which existing images it references, and generate a report on the likelihood that the image is truly original."

[0561] This system allows users to quickly and accurately determine the authenticity of images and take necessary measures in real time.

[0562] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0563] Step 1:

[0564] The user uses the device to select the image data they wish to analyze and upload it to the server. The input is the image data selected by the user operating the device, and the output is the image data sent to the server. Specifically, the user selects "Important Document.png" from the smartphone app and presses the upload button.

[0565] Step 2:

[0566] The server receives the image data sent by the user. The input is the image data sent in step 1, and the output is the image data saved in the server. Specifically, the server receives the HTTP request and saves the image file in a save directory.

[0567] Step 3:

[0568] Preprocessing is performed. The server resizes the received image data and removes noise. The input is the stored image data, and the output is the preprocessed image data. Specifically, the image is resized to 256x256 pixels and noise filtering is applied.

[0569] Step 4:

[0570] Extract feature vectors. The server generates feature vectors from the preprocessed image data using edge detection and color histograms. The input is the preprocessed image data, and the output is a feature vector. Specifically, it extracts edges from the image and quantifies the color distribution using a color histogram.

[0571] Step 5:

[0572] Determine the possibility of AI generation. The server inputs the feature vector into a machine learning model to determine whether the image was generated by AI. The input is the feature vector, and the output is a probability score of AI generation. Specifically, the feature vector is input into the TensorFlow model and the determination result is obtained.

[0573] Step 6:

[0574] The source of the image is identified. The server compares the feature vector with images in an existing database and identifies the most similar image. The input is the feature vector, and the output is the most similar existing image and its similarity. Specifically, the similarity is calculated using Sklearn's cosine_similarity function, and the index of the maximum value is obtained.

[0575] Step 7:

[0576] Calculate the reference ratio. The server calculates the reference ratio based on the similarity with the identified source image. The input is the similarity with the source image, and the output is the reference ratio. Specifically, the server expresses the reference ratio as a percentage based on the similarity score.

[0577] Step 8:

[0578] A report is generated and sent to the user. The server creates a detailed report based on the judgment results and reference ratio and sends it to the user's device. The input is the AI-generated likelihood score and reference ratio, and the output is report data. Specifically, the analysis results are summarized in text format and sent to the user via email or app notification.

[0579] This allows users to immediately check the analysis results and determine the authenticity of the image.

[0580] 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.

[0581] 1. System Configuration

[0582] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.It also combines an emotion engine that recognizes the user's emotions.

[0583] 2. Program Overview

[0584] The server performs a series of analytical processes on image data uploaded by users, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, analysis result report generation, and user emotion recognition using an emotion engine. The detailed operation of each step is explained below.

[0585] 3. User Operation

[0586] The user selects the image data they wish to analyze and uploads it to the server from their device. The device also captures the user's facial expressions and voice data for emotion recognition. User operation is performed via an intuitive and simple interface.

[0587] 4. Server Processing Details

[0588] Pretreatment

[0589] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[0590] Feature Extraction

[0591] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[0592] AI generation judgment

[0593] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[0594] Identifying the source of the misappropriation

[0595] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[0596] Calculating the Reference Ratio

[0597] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0598] 5. Emotion Engine Processing

[0599] Capturing Emotional Data

[0600] The device captures the user's facial and voice data and sends it to the emotion engine, which then analyzes the user's facial and voice data to recognize their emotional state.

[0601] Emotion analysis

[0602] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, and this emotion data is reflected in the analysis results.

[0603] 6. Generate and send reports

[0604] The server generates a detailed report based on the judgment results, reference ratio, and the emotion engine's recognition results. The report includes the possibility that the image was generated by AI, the detected source of misappropriation, the reference ratio, and a corresponding message based on the user's emotion. The report output format is adjusted taking this data into account.

[0605] Specific examples

[0606] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following process: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms. Then, use a machine learning model to determine whether the image is AI-generated and match it with existing images in the database to identify the source. Calculate the reference ratio and generate a detailed report. Finally, the report output format is adjusted based on the user's emotional state of excitement, and sent from the server to the user.

[0607] In this way, by combining the emotion engine, flexible report creation and notification based on the user's emotions becomes possible, making the way analysis results are received more user-friendly. This system is highly useful in a variety of fields.

[0608] The processing flow will be explained below.

[0609] Step 1: Upload user image

[0610] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data along with the user's facial expression and voice data to the server.

[0611] Step 2: The server receives the image data

[0612] The server receives the image data sent by the user, checks the format of the image data, and determines whether it can be analyzed.

[0613] Step 3: The server performs image preprocessing

[0614] The server performs preprocessing on the received image data. This includes image resizing and noise removal. Preprocessing is a necessary preparation to improve the accuracy of analysis.

[0615] Step 4: The server extracts the feature vectors

[0616] The server uses techniques such as edge detection and color histograms to extract feature vectors from the pre-processed images, which represent specific patterns or characteristics of the image as numerical data.

[0617] Step 5: The server determines the possibility of AI generation

[0618] The server uses the extracted feature vector to input data into a machine learning model, which evaluates the likelihood of AI generation and calculates a probability score. Based on this probability score, the server determines whether the image is AI-generated.

[0619] Step 6: The server identifies potential sources of misappropriation

[0620] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[0621] Step 7: The server calculates the referral rate

[0622] The server compares the uploaded image with the potential source image and calculates the reference ratio, which clarifies the extent to which the uploaded image references other images.

[0623] Step 8: The device captures emotion data

[0624] The device captures the user's facial expressions and voice data and sends it to the emotion engine, which allows the user's emotional state to be analyzed in real time.

[0625] Step 9: The server analyzes the emotion data

[0626] The server uses an emotion engine to analyze facial expressions and voice data sent from the device, and classifies the user's emotions into categories such as "happiness," "surprise," "sadness," and "anger."

[0627] Step 10: Server Generates Report

[0628] The server generates a detailed report based on the AI-generated judgment results, the reference ratio, and the emotion engine's recognition results, including the possibility that the image was generated by AI, the detected source of misappropriation, the reference ratio, and a corresponding message according to the user's emotion.

[0629] Step 11: Server sends report to user

[0630] The server sends the generated report to the user, who receives the report on their device and checks the analysis results and corresponding messages.

[0631] Example 2

[0632] 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."

[0633] Conventional image analysis systems can perform preprocessing of image data, extraction of feature vectors, determination of the possibility of AI generation, calculation of similarity, and calculation of reference ratios, but they lack the functionality to reflect the user's emotional state in the analysis results, which means that the user experience is not sufficiently improved.Another problem is that the reports generated based on the analysis results are uniform, making it impossible to respond flexibly to the user's emotions.

[0634] 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.

[0635] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for generating a report based on the determination result and the reference ratio and sending it to the user, means for receiving the user's facial expression and voice data and recognizing their emotional state, and means for adjusting the report output format based on the recognized emotional state. This makes it possible to generate reports that reflect the user's emotional state, making it more user-friendly to receive analysis results.

[0636] "Image data" means visual information represented in digital form.

[0637] "Preprocessing" refers to initial data processing operations performed to improve the accuracy of the analysis of image data.

[0638] A "feature vector" is a numerical representation of characteristics and patterns extracted from image data.

[0639] "AI-generated likelihood determination" is the process of using the extracted feature vectors to determine whether an image was generated by artificial intelligence.

[0640] "Calculation of similarity" is a process of numerically evaluating the similarity between image data and existing image data based on the feature vector.

[0641] The "reference ratio" is a ratio that quantitatively indicates to what extent the uploaded image references existing image data.

[0642] "Report generation" is the process of creating a detailed report based on the analysis results to provide to the user.

[0643] "User's facial expression and voice data" refers to data that digitally represents the user's facial expression and vocal content.

[0644] "Emotional state recognition" is the process of analyzing a user's facial expressions and voice data to determine the user's emotional state.

[0645] "Adjusting report output format" refers to flexibly changing the appearance and content of generated reports based on the perceived emotional state.

[0646] In this system, users upload image data to a server using their devices, and the server analyzes the image data to identify the possibility of AI generation, the reference source, and the reference ratio. Furthermore, it combines an emotion engine that recognizes the user's emotions.

[0647] 1. System Configuration

[0648] The system consists of the following main components:

[0649] Terminal

[0650] server

[0651] Emotion Engine

[0652] 2. User Operation

[0653] Users select the image data they wish to analyze and upload it to the server from their device. In addition to the image data, the device also captures the user's facial expressions and voice data and sends this data to the server. Users operate the system via an intuitive and simple interface.

[0654] 3. Image data processing details (server operation)

[0655] The server processes the received image data in the following steps:

[0656] 1. Pretreatment

[0657] The server resizes the images and removes noise, specifically using libraries such as OpenCV.

[0658] 2. Extracting feature vectors

[0659] Extract feature vectors from preprocessed image data using edge detection (e.g., Canny method) and color histograms. Use functions such as cv2.calcHist.

[0660] 3. AI-generated possibility assessment

[0661] The feature vector is fed into a machine learning model (e.g., a TensorFlow model) to assess the likelihood that the image was generated by AI.

[0662] 4. Identifying the source of the misappropriation

[0663] The feature vectors are matched with existing images in a database, and potential sources are identified based on similarity (e.g., cosine similarity).

[0664] 5. Calculation of Reference Ratio

[0665] The reference ratio is calculated for the original image, and the extent to which the existing image is referenced is quantitatively shown.

[0666] 4. Emotion engine processing (terminal and server operation)

[0667] 1. Capturing Emotional Data

[0668] The device captures the user's facial expressions and voice data and sends that data to the emotion engine.

[0669] 2. Emotion analysis

[0670] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, using facial recognition algorithms and voice emotion recognition models.

[0671] 5. Generate and send reports (server operations)

[0672] The server generates a detailed report based on the judgment results, reference ratios, and the emotion engine's recognition results, including:

[0673] The image may have been generated by AI

[0674] Identified potential sources of misappropriation

[0675] Reference ratio

[0676] Response messages based on user sentiment

[0677] This report is sent to the user, and the output format is adjusted according to the user's feelings.

[0678] Specific examples

[0679] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following processes: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms, and use a machine learning model to determine the possibility of AI generation. Match the image with existing images in the database to identify the source and calculate the reference ratio. Generate a detailed report, adjust the report output format based on the user's emotional state of excitement, and send it to the user from the server.

[0680] Example prompts for generative AI models

[0681] "Please determine whether the uploaded image data was generated by AI and analyze whether it has a specific source. Also, please determine the user's emotion and generate a report based on that. For example, the image file name is 'sunset.jpg' and the user's emotion is 'excited'."

[0682] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0683] Step 1: Upload image data

[0684] Input: An image file selected by the user on the device.

[0685] Processing: The user selects the image data they wish to analyze and uploads it from their device to the server.

[0686] Output: Image file uploaded to the server

[0687] Specific operation: The user selects the desired image file (e.g., "sunset.jpg") from the device's file system and clicks the "Upload" button, which sends the image data to the server.

[0688] Step 2: Capturing emotion data

[0689] Input: User facial and voice data

[0690] Processing: The device captures the user's facial expressions and voice and sends them to the server as emotional data.

[0691] Output: Facial expression and voice data sent to the server

[0692] Specific operation: The device's camera and microphone are used to record and photograph the user's face and voice in real time, and the recorded and photographed data is sent to a server via the network.

[0693] Step 3: Preprocessing the image data

[0694] Input: Image file uploaded to the server

[0695] Processing: The server performs preprocessing on the received image data.

[0696] Output: Preprocessed image data

[0697] Specific operation: The server uses a library (e.g. OpenCV) to resize the image data to a standard size and remove noise using filtering techniques.

[0698] Step 4: Extract feature vectors

[0699] Input: Preprocessed image data

[0700] Processing: The server extracts feature vectors from the preprocessed image data.

[0701] Output: Extracted feature vector

[0702] Specific operation: The server uses edge detection (e.g., Canny method) and color histogram (e.g., cv2.calcHist function) to extract image features as numerical data.

[0703] Step 5: Determine the possibility of AI generation

[0704] Input: Extracted feature vector

[0705] Processing: The server inputs the extracted feature vectors into a machine learning model to determine the likelihood that the image was generated by AI.

[0706] Output: Probability score indicating the likelihood of the AI ​​generation

[0707] What it does: The server uses a trained machine learning model (e.g., a TensorFlow model) to evaluate feature vectors and output an AI-generated probability score.

[0708] Step 6: Identify the source

[0709] Input: Extracted feature vector

[0710] Processing: The server matches the feature vector with existing images in a database to identify potential sources.

[0711] Output: Identified potential sources of misappropriation

[0712] Specific operation: The server queries the database for the feature vector and identifies the closest match based on a calculated similarity (e.g., cosine similarity).

[0713] Step 7: Calculate the reference ratio

[0714] Input: Uploaded image file, identified potential source

[0715] Processing: The server calculates the reference ratio for the identified source image.

[0716] Output: Reference ratio

[0717] Specific operation: The server calculates the reference ratio based on the similarity between the uploaded image and the original image, and quantitatively indicates the extent to which the image has been referenced.

[0718] Step 8: Analyze the sentiment data

[0719] Input: Facial and voice data sent to the server

[0720] Processing: The server's emotion engine analyzes the received facial and voice data to recognize the user's emotional state.

[0721] Output: User's emotional state

[0722] Specific operation: The server uses an emotion analysis algorithm (e.g., a facial recognition algorithm or a voice emotion recognition model) to analyze emotions from the user's facial expressions and voice.

[0723] Step 9: Generate reports

[0724] Input: Probability score indicating the likelihood of AI generation, identified potential sources of misappropriation, reference ratio, and user emotional state

[0725] Processing: The server generates a report based on the judgment results, reference ratio, and the emotion engine's recognition results.

[0726] Output: Detailed report for the user

[0727] Specific operation: The server organizes the analysis results and outputs a report in a format that is easy for the user to understand (e.g., PDF or HTML format). The report includes the possibility of the image being generated by AI, potential sources of reuse, the reference ratio, and a corresponding message based on the user's emotions.

[0728] Step 10: Submit the report

[0729] Input: Detailed reports for users

[0730] Processing: The server generates the report and sends it to the user.

[0731] Output: Report sent to user

[0732] What happens: The server sends the report to the user via email or a download link. The user can download the report by clicking the provided link.

[0733] (Application example 2)

[0734] 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."

[0735] There is a need to verify the originality of image materials used by advertising creators and optimize ad placement based on user emotions, but conventional systems have had difficulty doing this effectively.In addition, there was no system that could flexibly respond by taking user emotions into account when evaluating the possibility that an image was generated by AI or identifying the source of the image.

[0736] 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 receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for capturing user emotion data and adjusting the analysis results based on the emotion data, and means for generating a report based on the determination results, reference ratio, and emotion data and sending it to the user. This makes it possible to confirm the originality of image materials used by advertising creators, evaluate the possibility of AI generation, and identify sources of reuse, as well as to suggest optimal ad placement based on user emotion.

[0737] "Image data" is digital data of visual information that a user uploads to a server using a terminal.

[0738] "Preprocessing" refers to initial processing such as resizing and noise removal that is performed before analyzing image data.

[0739] A "feature vector" is a numerical representation of a particular pattern or characteristic of image data.

[0740] "Likelihood of AI Generation" is a probability score that assesses whether image data was generated by artificial intelligence.

[0741] "Similarity" is a numerical representation of the degree of match between feature vectors.

[0742] The "reference ratio" is the percentage that indicates the extent to which the uploaded image references an existing image.

[0743] "Emotion data" is information that indicates the emotional state of the user analyzed from facial expressions and voice data.

[0744] "Capture" refers to the device recording the user's facial expressions and voice and acquiring them as digital data.

[0745] The "analysis results" are a compilation of all the processing results from the server's preprocessing of image data, feature extraction, AI generation judgment, similarity calculation, reference ratio calculation, and emotion analysis.

[0746] A "report" is a detailed report generated based on the analysis results and emotion data and provided to the user.

[0747] 1. System Configuration

[0748] This invention is a system aimed at verifying the originality of image materials used by advertising creators and optimizing advertising placement based on user emotions. The system consists of three main elements: a server, a terminal, and a user. The server analyzes image data, the terminal captures image data and emotional data from the user, and the user uploads the image data.

[0749] 2. Program Overview

[0750] To implement this invention, the following hardware and software are used. The hardware requires a terminal equipped with a camera for capturing images and a microphone for capturing audio, and a server capable of high-speed computation. The software requires a library using OpenCV (cv2) to resize and denoise images, an engine using EmotionEngine to recognize user emotions, a module using FeatureExtractor to extract feature vectors, a module using AIDetector to determine the possibility of AI generation, a module using SimilaritySearch to identify the source of reuse, and a module using ReportGenerator to generate the final report.

[0751] 3. Processing Flow

[0752] The user selects the image data they wish to analyze and uploads it from their device to the server. The device simultaneously captures the user's facial expressions and voice data and transmits their emotional state to the server. The server first performs preprocessing on the received image data, including image resizing and noise removal. After the preprocessing is complete, feature vectors are extracted from the image data. Edge detection and color histograms are used to extract the feature vectors.

[0753] The server then uses the extracted feature vectors to assess the likelihood of AI generation. This is done using a machine learning model previously trained on a large amount of data. The result of the AI ​​generation assessment is a probability score indicating whether the image was generated by AI. The server then matches the feature vectors with existing images in its database to identify the source of the image. For each identified source image, it calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references existing images.

[0754] Furthermore, the emotion engine analyzes the user's emotional state and reflects this emotional data in the analysis results. Finally, the server generates a detailed report based on the analysis results, reference ratio, and emotion data and sends it to the user. The report includes the possibility that the image was generated by AI, the original image and reference ratio, and a corresponding message based on the user's emotion.

[0755] 4. Specific Examples

[0756] For example, consider the case where an ad creator uploads a new ad banner image "ad_banner.jpg" to the app and simultaneously captures user sentiment data (interestingness). The server receives this image, resizes it, and denoises it. Next, edge detection and color histograms are used to extract feature vectors. A machine learning model is then used to determine whether it is AI-generated and match it with existing images in the database to identify the source. The reference ratio is calculated and a detailed report is generated. Finally, a report containing optimal ad placement suggestions based on user sentiment is generated and sent to the creator.

[0757] Example prompt sentence:

[0758] "Upload your ad banner image and see how original and AI-generated it is."

[0759] "Please suggest the best ad placement if the user finds it interesting."

[0760] In this way, this system can verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose optimal ad placements based on user sentiment.

[0761] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0762] Step 1:

[0763] The user selects the image data they wish to analyze and uploads it from their device to the server. The input is the image data "ad_banner.jpg" selected by the user, and the output is the image data sent to the server. In this process, the device reads the image data and sends it to the specified server's upload endpoint.

[0764] Step 2:

[0765] The device captures the user's facial and voice data and sends the emotion data to the server. The input is the user's facial and voice data, and the output is the emotion data sent to the server. In this process, the device captures data in real time using a camera and microphone and sends it to the server in a format for emotion analysis.

[0766] Step 3:

[0767] The server performs preprocessing on the received image data. The input is the image data "ad_banner.jpg" sent to the server, and the output is the image data after resizing and noise removal. In this process, OpenCV (cv2) is used to resize the image to 256x256 pixels and apply a noise removal algorithm.

[0768] Step 4:

[0769] The server extracts feature vectors from the preprocessed image data. The input is the preprocessed image data, and the output is a feature vector. In this process, FeatureExtractor is used to apply edge detection and color histogram techniques to extract image characteristics as numerical data.

[0770] Step 5:

[0771] The server uses the extracted feature vector to determine the possibility of AI generation. The input is the extracted feature vector, and the output is a probability score indicating the possibility of AI generation. In this process, the feature vector is input into the pre-trained AIDetector machine learning model to evaluate the possibility of AI generation.

[0772] Step 6:

[0773] The server compares the feature vector with existing images in the database and calculates the similarity. The input is the extracted feature vector, and the output is the similarity score with the existing image and the candidate image to be reused. In this process, SimilaritySearch is used to compare the feature vector with images in the database and identify the most similar image.

[0774] Step 7:

[0775] The server calculates the reference ratio. The input is the similarity score with existing images, and the output is the reference ratio. This process quantitatively calculates how much the uploaded image references existing images based on the similarity score.

[0776] Step 8:

[0777] The server analyzes the user's emotional data and adjusts the analysis results based on that emotional state. The input is the emotional data, and the output is the adjusted analysis results. In this process, the Emotion Engine is used to recognize the user's emotional state, and the analysis results reflect responses that match those emotions.

[0778] Step 9:

[0779] The server generates a report based on the judgment results, reference ratios, and emotion data and sends it to the user. The inputs are the judgment results, reference ratios, and emotion data, and the output is the generated report. In this process, the ReportGenerator aggregates the analysis results and creates and notifies the user of the report in an easy-to-understand format.

[0780] Through each of the above steps, this system is able to verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose appropriate ad placements based on user sentiment.

[0781] 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.

[0782] 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.

[0783] 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.

[0784] [Third embodiment]

[0785] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0786] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0787] 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).

[0788] 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.

[0789] 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.

[0790] 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).

[0791] 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.

[0792] 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.

[0793] 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.

[0794] 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.

[0795] 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.

[0796] 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."

[0797] 1. System Configuration

[0798] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[0799] 2. Program Overview

[0800] The server performs a series of analytical processes on the image data uploaded by the user, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, and analysis result report generation. The detailed operation of each step is explained below.

[0801] 3. User Operation

[0802] Users select the image data they wish to analyze and upload it to the server from their device. User operations are performed via an intuitive and simple interface.

[0803] 4. Server Processing Details

[0804] Pretreatment

[0805] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[0806] Feature Extraction

[0807] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[0808] AI generation judgment

[0809] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[0810] Identifying the source of the misappropriation

[0811] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[0812] Calculating the Reference Ratio

[0813] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0814] 5. Generate and send the report

[0815] The server generates a detailed report based on the results of the judgment and the analysis of the reference ratio. This report includes data on the possibility of AI generation, the identification of the source of misuse, and the reference ratio. The generated report is sent from the server to the user, who can view it on their device.

[0816] Specific examples

[0817] For example, if a user uploads an image called "sunset.jpg," the server performs the following processes in order: Receive "sunset.jpg" and perform appropriate resizing and noise reduction. Next, extract feature vectors using edge detection and color histograms. After that, use a machine learning model to determine whether it is AI-generated and compare it with existing images in the database to identify the source. Finally, calculate the reference ratio between "sunset.jpg" and the most similar image, and send a report summarizing all these results to the user.

[0818] In this way, the present invention allows users to easily and efficiently obtain the origin and reference information of an image, making it easier to verify the authenticity of the image and ensure its originality. This system is highly useful in a variety of application fields.

[0819] The processing flow will be explained below.

[0820] Step 1: Upload user image

[0821] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data to the server.

[0822] Step 2: The server receives the image data

[0823] The server receives the image data sent by the user and checks the format of the image data to see if there are any problems.

[0824] Step 3: The server performs image preprocessing

[0825] The server performs preprocessing on the received image data, including image resizing and noise removal, to prepare the image for further analysis.

[0826] Step 4: The server extracts the feature vectors

[0827] The server extracts feature vectors from the preprocessed images, using techniques such as edge detection and color histograms to extract specific patterns and characteristics of the image as numerical data.

[0828] Step 5: The server determines the possibility of AI generation

[0829] The server inputs the extracted feature vector into a machine learning model, which evaluates the likelihood of AI generation and generates a probability score. The server then uses this probability score to determine whether the image was AI-generated.

[0830] Step 6: The server identifies potential sources of misappropriation

[0831] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[0832] Step 7: The server calculates the referral rate

[0833] The server calculates the reference ratio using the uploaded image and the candidate image for reuse. The reference ratio is a numerical value that indicates the degree to which a specific image references other images.

[0834] Step 8: Server Generates Report

[0835] The server generates a detailed report based on the AI-generated image judgment results and the analysis of the reference ratio, including the probability that the image was generated by AI, the detected potential sources of misappropriation, and the reference ratio.

[0836] Step 9: Server sends report to user

[0837] The server sends the generated report to the user, who receives the report on their device and checks the analysis results.

[0838] Example 1

[0839] 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."

[0840] Conventional image analysis systems have had difficulty efficiently determining whether image data uploaded by users is AI-generated, identifying the source of misuse, and calculating the reference ratio. In particular, it has been difficult to automatically and accurately perform a series of processes, from highly accurate preprocessing, feature extraction, AI generation determination, reference ratio calculation, and report generation. The purpose of this invention is to address these challenges and provide a system that allows users to easily verify the authenticity and originality of image data.

[0841] 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.

[0842] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This makes it possible to accurately and quickly determine the possibility of AI generation of image data uploaded by a user and identify the source of the misuse.

[0843] "Image data" refers to a digital file format that electronically stores visual information.

[0844] "Preprocessing" refers to initial processing such as resizing and noise removal performed on image data to improve the accuracy of analysis.

[0845] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[0846] "AI generation determination" refers to the process of using a machine learning model to determine whether an image was generated by artificial intelligence.

[0847] "Similarity" is a value calculated by calculating the degree of match between the feature vectors of two images.

[0848] The "reference ratio" is a quantitative ratio that indicates the extent to which an uploaded image references an existing image.

[0849] A "report" is a document summarizing the results of analysis such as the possibility of AI generation, identification of the source of misuse, and reference ratio.

[0850] A "database" is a collection of information that registers and manages a large number of known image data.

[0851] A "machine learning model" is an algorithm that is trained on a large amount of data in advance and then uses the results to analyze new data.

[0852] "Edge detection" is an image processing technique that identifies edges and boundaries within an image.

[0853] A "color histogram" is statistical data that represents the distribution of colors contained in an image.

[0854] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[0855] Hardware and software used

[0856] The following hardware and software are used to implement the present invention.

[0857] Hardware: Server machine equipped with high-speed CPU and GPU

[0858] Software: Image processing libraries (e.g., OpenCV), machine learning libraries (e.g., TensorFlow, PyTorch), database management systems (e.g., MySQL)

[0859] Example of operation

[0860] For example, if a user wants to upload an image called "sunset.jpg", the process is as follows:

[0861] 1. User operations

[0862] The user selects an image file "sunset.jpg" and uploads it to the server via the terminal. An example prompt is as follows:

[0863] Input image: sunset.jpg

[0864] 2. Server Reception

[0865] The server receives the uploaded "sunset.jpg" file and saves it in the specified directory. A message indicating that the file has been received is recorded in the server log.

[0866] 3. Pretreatment

[0867] The server performs preprocessing on the received image files, including image resizing and noise removal. Specifically, the server uses OpenCV to resize the image to a specific size and remove noise.

[0868] 4. Feature Extraction

[0869] After preprocessing, the server extracts feature vectors from the images using edge detection and color histograms, which represent specific patterns and characteristics of the image as numerical data.

[0870] 5. AI generation judgment

[0871] The extracted feature vectors are input into a machine learning model to determine whether the image is AI-generated or not, using a model trained in TensorFlow or PyTorch to obtain a probability score for the likelihood of AI generation.

[0872] 6. Identifying the source of the misappropriation

[0873] The server compares the feature vectors with existing images in a database, which contains a large number of known image data, and identifies the closest candidates based on the calculated similarity.

[0874] 7. Calculation of Reference Ratio

[0875] For each identified source image, the server calculates the reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0876] 8. Generate and send reports

[0877] The server compiles information on the possibility of AI generation, the identification of the source of misuse, and the reference ratio to generate a detailed report. This report is generated in PDF or HTML format and sent to the user, who can then view the analysis results on their device.

[0878] The present invention provides a system that can accurately identify the possibility that image data uploaded by users is AI-generated and the source of the misuse, allowing users to easily verify the authenticity and originality of images.

[0879] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0880] Step 1: User selects and uploads image data

[0881] Users select the image data they wish to analyze from their own device. For example, they select an image file called "sunset.jpg" and upload it to the server via an intuitive interface. They select the "sunset.jpg" file as input and click the "Upload" button. The image data is sent to the server as output.

[0882] Step 2: The server receives the image data.

[0883] The server receives the "sunset.jpg" file uploaded by the user. As input, the server receives the image data sent from the user's device and saves it in the specified directory. Specifically, a message indicating completion of reception is recorded in the server's log. As output, the image data is saved on the server in a state where it can be preprocessed.

[0884] Step 3: Server preprocessing

[0885] The server performs preprocessing on the received image data. It reads the "sunset.jpg" file as input and performs the following operations: it resizes the image to a specific size using the OpenCV library (cv2.resize(image, (256, 256))) and removes noise (cv2.fastNlMeansDenoisingColored(image, None, 10, 10, 7, 21)). The output is the preprocessed image data.

[0886] Step 4: The server performs feature extraction

[0887] After preprocessing, the server extracts a feature vector from the image data. Using the preprocessed image data as input, it performs the following operations: edge detection (cv2.Canny(image, 100, 200)) and color histogram (cv2.calcHist([image], [0, 1, 2], None, [8, 8, 8], [0, 256, 0, 256, 0, 256])). The output is the feature vector of the image.

[0888] Step 5: The server makes the AI ​​generation decision

[0889] The server inputs the extracted feature vector into a machine learning model to determine whether the image was generated by AI. The feature vector is input into the machine learning model as input, and the following process is performed: Specifically, the feature vector is analyzed using a pre-trained model (e.g., TensorFlow or PyTorch) (model(features)). The output is a probability score for the likelihood of AI generation.

[0890] Step 6: The server identifies the source of the stolen data

[0891] The server compares the extracted feature vector with existing images in the database. As input, it uses the feature vector to calculate the similarity with existing images in the database. Specifically, it performs the similarity calculation and identifies the candidate source with the closest match (compare_with_database(features)). As output, it identifies the most similar existing image.

[0892] Step 7: The server calculates the referral rate.

[0893] Based on the identified source image, the server calculates the reference ratio. Using the feature vectors of the identified source image and the uploaded image data as input, the server executes the following process. Specifically, it calculates the reference ratio with the source image (calculate_reference_ratio(features, best_match_features)). The output shows the reference ratio quantitatively.

[0894] Step 8: The server generates the report and sends it to the user

[0895] The server generates a detailed report based on the analysis results and sends it to the user. It uses the AI-generated probability score, misappropriation source identification, and reference ratio data as input and performs the following process: Specifically, it generates a report in PDF or HTML format (generate_report(probability_score, best_match_details, reference_ratio)) and sends it to the user (send_report_to_user(report)). The analysis results are provided to the user as output.

[0896] (Application example 1)

[0897] 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."

[0898] In modern society, the ease with which counterfeit and inappropriate images are spread over the Internet is increasing, and these false information and inappropriate images can cause serious problems for individuals, companies, and organizations. Therefore, there is a need to quickly and accurately determine the authenticity of images and take appropriate countermeasures as necessary. However, with current technology, determining the authenticity of images and detecting counterfeits often requires time and effort, making real-time detection difficult. This presents a challenge in terms of making it difficult to respond quickly and effectively.

[0899] 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.

[0900] In this invention, the server includes means for receiving and preprocessing image data, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for determining whether the image data is forged, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This enables users to quickly and accurately determine whether images on the Internet are forged and take appropriate measures in real time based on the results.

[0901] "Image data" is visual information represented in digital form.

[0902] "Preprocessing" refers to preparatory work to improve the accuracy of analysis, such as resizing image data and removing noise.

[0903] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[0904] "Likelihood of AI Generation" is a probability score that indicates whether an image was generated by artificial intelligence.

[0905] The "similarity" is a value calculated by using feature vectors to determine the similarity between different images.

[0906] The "reference ratio" is a quantitative index that indicates the degree to which an image references an existing image.

[0907] "Counterfeit determination" is the process of assessing whether an image is counterfeit.

[0908] A "report" is a document summarizing the analysis results and is information provided to the user.

[0909] A "server" is a computer system that processes data on a network.

[0910] A "user" is an individual or organization that uses the system to upload image data and receive analysis results.

[0911] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio. This system is particularly useful for detecting counterfeit images in security services.

[0912] 1. System Configuration

[0913] The present invention mainly comprises the following elements:

[0914] Hardware

[0915] Device: A device that takes and uploads images, such as a smartphone or smart glasses.

[0916] Server: A computer system that receives image data and performs analysis and processing.

[0917] software

[0918] Python: A programming language for writing the overall program logic.

[0919] TensorFlow: A library for extracting feature vectors and running machine learning models to determine AI-generated possibilities.

[0920] Sklearn: A library for computing similarities between feature vectors.

[0921] 2. Specific processing flow

[0922] User operations

[0923] Users use the terminal to select the image data they wish to analyze and upload it to the server. The terminal has an intuitive and simple interface, so users can operate it without any special skills.

[0924] Server Processing

[0925] 1. Pretreatment

[0926] The server receives the image data sent by the user and performs preprocessing such as resizing and noise removal on the received image data.

[0927] 2. Extracting feature vectors

[0928] After preprocessing is complete, feature vectors are extracted from the image data using techniques such as edge detection and color histograms.

[0929] 3. AI-generated judgment

[0930] The extracted feature vector is input into a machine learning model to determine whether the image was generated by AI. The machine learning model is built using TensorFlow and is pre-trained with a large amount of data.

[0931] 4. Identifying the source of the misappropriation

[0932] The feature vector is matched with existing images in a database to identify the most similar images. Similarity is calculated using the cosine_similarity function in Sklearn.

[0933] 5. Calculation of reference ratio

[0934] The reference ratio is calculated based on the similarity between the original image and the uploaded image.

[0935] 6. Report Generation and Notifications

[0936] A detailed report is generated based on the assessment results and reference ratio and sent to the user, who can then check the report on their device and take any necessary measures based on the information.

[0937] Specific examples

[0938] If a user takes an image titled "Important Document.png" during an important company meeting and wants to check whether the image may be forged, they can use this system. The user uploads the image to the server, which performs a series of analysis processes and then provides the user with a determination of whether the image is forged.

[0939] Prompt Sentence Examples

[0940] "Determine if this image was generated by AI, identify which existing images it references, and generate a report on the likelihood that the image is truly original."

[0941] This system allows users to quickly and accurately determine the authenticity of images and take necessary measures in real time.

[0942] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0943] Step 1:

[0944] The user uses the device to select the image data they wish to analyze and upload it to the server. The input is the image data selected by the user operating the device, and the output is the image data sent to the server. Specifically, the user selects "Important Document.png" from the smartphone app and presses the upload button.

[0945] Step 2:

[0946] The server receives the image data sent by the user. The input is the image data sent in step 1, and the output is the image data saved in the server. Specifically, the server receives the HTTP request and saves the image file in a save directory.

[0947] Step 3:

[0948] Preprocessing is performed. The server resizes the received image data and removes noise. The input is the stored image data, and the output is the preprocessed image data. Specifically, the image is resized to 256x256 pixels and noise filtering is applied.

[0949] Step 4:

[0950] Extract feature vectors. The server generates feature vectors from the preprocessed image data using edge detection and color histograms. The input is the preprocessed image data, and the output is a feature vector. Specifically, it extracts edges from the image and quantifies the color distribution using a color histogram.

[0951] Step 5:

[0952] Determine the possibility of AI generation. The server inputs the feature vector into a machine learning model to determine whether the image was generated by AI. The input is the feature vector, and the output is a probability score of AI generation. Specifically, the feature vector is input into the TensorFlow model and the determination result is obtained.

[0953] Step 6:

[0954] The source of the image is identified. The server compares the feature vector with images in an existing database and identifies the most similar image. The input is the feature vector, and the output is the most similar existing image and its similarity. Specifically, the similarity is calculated using Sklearn's cosine_similarity function, and the index of the maximum value is obtained.

[0955] Step 7:

[0956] Calculate the reference ratio. The server calculates the reference ratio based on the similarity with the identified source image. The input is the similarity with the source image, and the output is the reference ratio. Specifically, the server expresses the reference ratio as a percentage based on the similarity score.

[0957] Step 8:

[0958] A report is generated and sent to the user. The server creates a detailed report based on the judgment results and reference ratio and sends it to the user's device. The input is the AI-generated likelihood score and reference ratio, and the output is report data. Specifically, the analysis results are summarized in text format and sent to the user via email or app notification.

[0959] This allows users to immediately check the analysis results and determine the authenticity of the image.

[0960] 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.

[0961] 1. System Configuration

[0962] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.It also combines an emotion engine that recognizes the user's emotions.

[0963] 2. Program Overview

[0964] The server performs a series of analytical processes on image data uploaded by users, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, analysis result report generation, and user emotion recognition using an emotion engine. The detailed operation of each step is explained below.

[0965] 3. User Operation

[0966] The user selects the image data they wish to analyze and uploads it to the server from their device. The device also captures the user's facial expressions and voice data for emotion recognition. User operation is performed via an intuitive and simple interface.

[0967] 4. Server Processing Details

[0968] Pretreatment

[0969] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[0970] Feature Extraction

[0971] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[0972] AI generation judgment

[0973] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[0974] Identifying the source of the misappropriation

[0975] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[0976] Calculating the Reference Ratio

[0977] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[0978] 5. Emotion Engine Processing

[0979] Capturing Emotional Data

[0980] The device captures the user's facial and voice data and sends it to the emotion engine, which then analyzes the user's facial and voice data to recognize their emotional state.

[0981] Emotion analysis

[0982] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, and this emotion data is reflected in the analysis results.

[0983] 6. Generate and send reports

[0984] The server generates a detailed report based on the judgment results, reference ratio, and the emotion engine's recognition results. The report includes the possibility that the image was generated by AI, the detected source of misappropriation, the reference ratio, and a corresponding message based on the user's emotion. The report output format is adjusted taking this data into account.

[0985] Specific examples

[0986] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following process: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms. Then, use a machine learning model to determine whether the image is AI-generated and match it with existing images in the database to identify the source. Calculate the reference ratio and generate a detailed report. Finally, the report output format is adjusted based on the user's emotional state of excitement, and sent from the server to the user.

[0987] In this way, by combining the emotion engine, flexible report creation and notification based on the user's emotions becomes possible, making the way analysis results are received more user-friendly. This system is highly useful in a variety of fields.

[0988] The processing flow will be explained below.

[0989] Step 1: Upload user image

[0990] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data along with the user's facial expression and voice data to the server.

[0991] Step 2: The server receives the image data

[0992] The server receives the image data sent by the user, checks the format of the image data, and determines whether it can be analyzed.

[0993] Step 3: The server performs image preprocessing

[0994] The server performs preprocessing on the received image data. This includes image resizing and noise removal. Preprocessing is a necessary preparation to improve the accuracy of analysis.

[0995] Step 4: The server extracts the feature vectors

[0996] The server uses techniques such as edge detection and color histograms to extract feature vectors from the pre-processed images, which represent specific patterns or characteristics of the image as numerical data.

[0997] Step 5: The server determines the possibility of AI generation

[0998] The server uses the extracted feature vector to input data into a machine learning model, which evaluates the likelihood of AI generation and calculates a probability score. Based on this probability score, the server determines whether the image is AI-generated.

[0999] Step 6: The server identifies potential sources of misappropriation

[1000] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[1001] Step 7: The server calculates the referral rate

[1002] The server compares the uploaded image with the potential source image and calculates the reference ratio, which clarifies the extent to which the uploaded image references other images.

[1003] Step 8: The device captures emotion data

[1004] The device captures the user's facial expressions and voice data and sends it to the emotion engine, which allows the user's emotional state to be analyzed in real time.

[1005] Step 9: The server analyzes the emotion data

[1006] The server uses an emotion engine to analyze facial expressions and voice data sent from the device, and classifies the user's emotions into categories such as "happiness," "surprise," "sadness," and "anger."

[1007] Step 10: Server Generates Report

[1008] The server generates a detailed report based on the AI-generated judgment results, reference ratio, and the emotion engine's recognition results, including the possibility that the image was generated by AI, the detected potential sources of misappropriation, the reference ratio, and a corresponding message according to the user's emotion.

[1009] Step 11: Server sends report to user

[1010] The server sends the generated report to the user, who receives the report on their device and checks the analysis results and corresponding messages.

[1011] Example 2

[1012] 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."

[1013] Conventional image analysis systems can perform preprocessing of image data, extraction of feature vectors, determination of the possibility of AI generation, calculation of similarity, and calculation of reference ratios, but they lack the functionality to reflect the user's emotional state in the analysis results, which means that the user experience is not sufficiently improved.Another problem is that the reports generated based on the analysis results are uniform, making it impossible to respond flexibly to the user's emotions.

[1014] 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.

[1015] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for generating a report based on the determination result and the reference ratio and sending it to the user, means for receiving the user's facial expression and voice data and recognizing their emotional state, and means for adjusting the report output format based on the recognized emotional state. This makes it possible to generate reports that reflect the user's emotional state, making it more user-friendly to receive analysis results.

[1016] "Image data" means visual information represented in digital form.

[1017] "Preprocessing" refers to initial data processing operations performed to improve the accuracy of the analysis of image data.

[1018] A "feature vector" is a numerical representation of characteristics and patterns extracted from image data.

[1019] "AI-generated likelihood determination" is the process of using the extracted feature vectors to determine whether an image was generated by artificial intelligence.

[1020] "Calculation of similarity" is a process of numerically evaluating the similarity between image data and existing image data based on the feature vector.

[1021] The "reference ratio" is a ratio that quantitatively indicates to what extent the uploaded image references existing image data.

[1022] "Report generation" is the process of creating a detailed report based on the analysis results to provide to the user.

[1023] "User's facial expression and voice data" refers to data that digitally represents the user's facial expression and vocal content.

[1024] "Emotional state recognition" is the process of analyzing a user's facial expressions and voice data to determine the user's emotional state.

[1025] "Adjusting report output format" refers to flexibly changing the appearance and content of generated reports based on the perceived emotional state.

[1026] In this system, users upload image data to a server using their devices, and the server analyzes the image data to identify the possibility of AI generation, the reference source, and the reference ratio. Furthermore, it combines an emotion engine that recognizes the user's emotions.

[1027] 1. System Configuration

[1028] The system consists of the following main components:

[1029] Terminal

[1030] server

[1031] Emotion Engine

[1032] 2. User Operation

[1033] Users select the image data they wish to analyze and upload it to the server from their device. In addition to the image data, the device also captures the user's facial expressions and voice data and sends this data to the server. Users operate the system via an intuitive and simple interface.

[1034] 3. Image data processing details (server operation)

[1035] The server processes the received image data in the following steps:

[1036] 1. Pretreatment

[1037] The server resizes the images and removes noise, specifically using libraries such as OpenCV.

[1038] 2. Extracting feature vectors

[1039] Extract feature vectors from preprocessed image data using edge detection (e.g., Canny method) and color histograms. Use functions such as cv2.calcHist.

[1040] 3. AI-generated possibility assessment

[1041] The feature vector is fed into a machine learning model (e.g., a TensorFlow model) to assess the likelihood that the image was generated by AI.

[1042] 4. Identifying the source of the misappropriation

[1043] The feature vectors are matched with existing images in a database, and potential sources are identified based on similarity (e.g., cosine similarity).

[1044] 5. Calculation of Reference Ratio

[1045] The reference ratio is calculated for the original image, and the extent to which the existing image is referenced is quantitatively shown.

[1046] 4. Emotion engine processing (terminal and server operation)

[1047] 1. Capturing Emotional Data

[1048] The device captures the user's facial expressions and voice data and sends that data to the emotion engine.

[1049] 2. Emotion analysis

[1050] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, using facial recognition algorithms and voice emotion recognition models.

[1051] 5. Generate and send reports (server operations)

[1052] The server generates a detailed report based on the judgment results, reference ratios, and the emotion engine's recognition results, including:

[1053] The image may have been generated by AI

[1054] Identified potential sources of misappropriation

[1055] Reference ratio

[1056] Response messages based on user sentiment

[1057] This report is sent to the user, and the output format is adjusted according to the user's feelings.

[1058] Specific examples

[1059] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following processes: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms, and use a machine learning model to determine the possibility of AI generation. Match the image with existing images in the database to identify the source and calculate the reference ratio. Generate a detailed report, adjust the report output format based on the user's emotional state of excitement, and send it to the user from the server.

[1060] Example prompts for generative AI models

[1061] "Please determine whether the uploaded image data was generated by AI and analyze whether it has a specific source. Also, please determine the user's emotion and generate a report based on that. For example, the image file name is 'sunset.jpg' and the user's emotion is 'excited'."

[1062] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1063] Step 1: Upload image data

[1064] Input: An image file selected by the user on the device.

[1065] Processing: The user selects the image data they wish to analyze and uploads it from their device to the server.

[1066] Output: Image file uploaded to the server

[1067] Specific operation: The user selects the desired image file (e.g., "sunset.jpg") from the device's file system and clicks the "Upload" button, which sends the image data to the server.

[1068] Step 2: Capturing emotion data

[1069] Input: User facial and voice data

[1070] Processing: The device captures the user's facial expressions and voice and sends them to the server as emotional data.

[1071] Output: Facial expression and voice data sent to the server

[1072] Specific operation: The device's camera and microphone are used to record and photograph the user's face and voice in real time, and the recorded and photographed data is sent to a server via the network.

[1073] Step 3: Preprocessing the image data

[1074] Input: Image file uploaded to the server

[1075] Processing: The server performs preprocessing on the received image data.

[1076] Output: Preprocessed image data

[1077] Specific operation: The server uses a library (e.g. OpenCV) to resize the image data to a standard size and remove noise using filtering techniques.

[1078] Step 4: Extract feature vectors

[1079] Input: Preprocessed image data

[1080] Processing: The server extracts feature vectors from the preprocessed image data.

[1081] Output: Extracted feature vector

[1082] Specific operation: The server uses edge detection (e.g., Canny method) and color histogram (e.g., cv2.calcHist function) to extract image features as numerical data.

[1083] Step 5: Determine the possibility of AI generation

[1084] Input: Extracted feature vector

[1085] Processing: The server inputs the extracted feature vectors into a machine learning model to determine the likelihood that the image was generated by AI.

[1086] Output: Probability score indicating the likelihood of the AI ​​generation

[1087] What it does: The server uses a trained machine learning model (e.g., a TensorFlow model) to evaluate feature vectors and output an AI-generated probability score.

[1088] Step 6: Identify the source

[1089] Input: Extracted feature vector

[1090] Processing: The server matches the feature vector with existing images in a database to identify potential sources.

[1091] Output: Identified potential sources of misappropriation

[1092] Specific operation: The server queries the database for the feature vector and identifies the closest match based on a calculated similarity (e.g., cosine similarity).

[1093] Step 7: Calculate the reference ratio

[1094] Input: Uploaded image file, identified potential source

[1095] Processing: The server calculates the reference ratio for the identified source image.

[1096] Output: Reference ratio

[1097] Specific operation: The server calculates the reference ratio based on the similarity between the uploaded image and the original image, and quantitatively indicates the extent to which the image has been referenced.

[1098] Step 8: Analyze the sentiment data

[1099] Input: Facial and voice data sent to the server

[1100] Processing: The server's emotion engine analyzes the received facial and voice data to recognize the user's emotional state.

[1101] Output: User's emotional state

[1102] Specific operation: The server uses an emotion analysis algorithm (e.g., a facial recognition algorithm or a voice emotion recognition model) to analyze emotions from the user's facial expressions and voice.

[1103] Step 9: Generate reports

[1104] Input: Probability score indicating the likelihood of AI generation, identified potential sources of misappropriation, reference ratio, and user emotional state

[1105] Processing: The server generates a report based on the judgment results, reference ratio, and the emotion engine's recognition results.

[1106] Output: Detailed report for the user

[1107] Specific operation: The server organizes the analysis results and outputs a report in a format that is easy for the user to understand (e.g., PDF or HTML format). The report includes the possibility of the image being generated by AI, potential sources of reuse, the reference ratio, and a corresponding message based on the user's emotions.

[1108] Step 10: Submit the report

[1109] Input: Detailed reports for users

[1110] Processing: The server generates the report and sends it to the user.

[1111] Output: Report sent to user

[1112] What happens: The server sends the report to the user via email or a download link. The user can download the report by clicking the provided link.

[1113] (Application example 2)

[1114] 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."

[1115] There is a need to verify the originality of image materials used by advertising creators and optimize ad placement based on user emotions, but conventional systems have had difficulty doing this effectively.In addition, there was no system that could flexibly respond by taking user emotions into account when evaluating the possibility that an image was generated by AI or identifying the source of the image.

[1116] 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 receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for capturing user emotion data and adjusting the analysis results based on the emotion data, and means for generating a report based on the determination results, reference ratio, and emotion data and sending it to the user. This makes it possible to confirm the originality of image materials used by advertising creators, evaluate the possibility of AI generation, and identify sources of reuse, as well as to suggest optimal ad placement based on user emotion.

[1117] "Image data" is digital data of visual information that a user uploads to a server using a terminal.

[1118] "Preprocessing" refers to initial processing such as resizing and noise removal that is performed before analyzing image data.

[1119] A "feature vector" is a numerical representation of a particular pattern or characteristic of image data.

[1120] "Likelihood of AI Generation" is a probability score that assesses whether image data was generated by artificial intelligence.

[1121] "Similarity" is a numerical representation of the degree of match between feature vectors.

[1122] The "reference ratio" is the percentage that indicates the extent to which the uploaded image references an existing image.

[1123] "Emotion data" is information that indicates the emotional state of the user analyzed from facial expressions and voice data.

[1124] "Capture" refers to the device recording the user's facial expressions and voice and acquiring them as digital data.

[1125] The "analysis results" are a compilation of all the processing results from the server's preprocessing of image data, feature extraction, AI generation judgment, similarity calculation, reference ratio calculation, and emotion analysis.

[1126] A "report" is a detailed report generated based on the analysis results and emotion data and provided to the user.

[1127] 1. System Configuration

[1128] This invention is a system aimed at verifying the originality of image materials used by advertising creators and optimizing advertising placement based on user emotions. The system consists of three main elements: a server, a terminal, and a user. The server analyzes image data, the terminal captures image data and emotional data from the user, and the user uploads the image data.

[1129] 2. Program Overview

[1130] To implement this invention, the following hardware and software are used. The hardware requires a terminal equipped with a camera for capturing images and a microphone for capturing audio, and a server capable of high-speed computation. The software requires a library using OpenCV (cv2) to resize and denoise images, an engine using EmotionEngine to recognize user emotions, a module using FeatureExtractor to extract feature vectors, a module using AIDetector to determine the possibility of AI generation, a module using SimilaritySearch to identify the source of reuse, and a module using ReportGenerator to generate the final report.

[1131] 3. Processing Flow

[1132] The user selects the image data they wish to analyze and uploads it from their device to the server. The device simultaneously captures the user's facial expressions and voice data and transmits their emotional state to the server. The server first performs preprocessing on the received image data, including image resizing and noise removal. After the preprocessing is complete, feature vectors are extracted from the image data. Edge detection and color histograms are used to extract the feature vectors.

[1133] The server then uses the extracted feature vectors to assess the likelihood of AI generation. This is done using a machine learning model previously trained on a large amount of data. The result of the AI ​​generation assessment is a probability score indicating whether the image was generated by AI. The server then matches the feature vectors with existing images in its database to identify the source of the image. For each identified source image, it calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references existing images.

[1134] Furthermore, the emotion engine analyzes the user's emotional state and reflects this emotional data in the analysis results. Finally, the server generates a detailed report based on the analysis results, reference ratio, and emotion data and sends it to the user. The report includes the possibility that the image was generated by AI, the original image and reference ratio, and a corresponding message based on the user's emotion.

[1135] 4. Specific Examples

[1136] For example, consider the case where an ad creator uploads a new ad banner image "ad_banner.jpg" to the app and simultaneously captures user sentiment data (interestingness). The server receives this image, resizes it, and denoises it. Next, edge detection and color histograms are used to extract feature vectors. A machine learning model is then used to determine whether it is AI-generated and match it with existing images in the database to identify the source. The reference ratio is calculated and a detailed report is generated. Finally, a report containing optimal ad placement suggestions based on user sentiment is generated and sent to the creator.

[1137] Example prompt sentence:

[1138] "Upload your ad banner image and see how original and AI-generated it is."

[1139] "Please suggest the best ad placement if the user finds it interesting."

[1140] In this way, this system can verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose optimal ad placements based on user sentiment.

[1141] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1142] Step 1:

[1143] The user selects the image data they wish to analyze and uploads it from their device to the server. The input is the image data "ad_banner.jpg" selected by the user, and the output is the image data sent to the server. In this process, the device reads the image data and sends it to the specified server's upload endpoint.

[1144] Step 2:

[1145] The device captures the user's facial and voice data and sends the emotion data to the server. The input is the user's facial and voice data, and the output is the emotion data sent to the server. In this process, the device captures data in real time using a camera and microphone and sends it to the server in a format for emotion analysis.

[1146] Step 3:

[1147] The server performs preprocessing on the received image data. The input is the image data "ad_banner.jpg" sent to the server, and the output is the image data after resizing and noise removal. In this process, OpenCV (cv2) is used to resize the image to 256x256 pixels and apply a noise removal algorithm.

[1148] Step 4:

[1149] The server extracts feature vectors from the preprocessed image data. The input is the preprocessed image data, and the output is a feature vector. In this process, FeatureExtractor is used to apply edge detection and color histogram techniques to extract image characteristics as numerical data.

[1150] Step 5:

[1151] The server uses the extracted feature vector to determine the possibility of AI generation. The input is the extracted feature vector, and the output is a probability score indicating the possibility of AI generation. In this process, the feature vector is input into the pre-trained AIDetector machine learning model to evaluate the possibility of AI generation.

[1152] Step 6:

[1153] The server compares the feature vector with existing images in the database and calculates the similarity. The input is the extracted feature vector, and the output is the similarity score with the existing image and the candidate image to be reused. In this process, SimilaritySearch is used to compare the feature vector with images in the database and identify the most similar image.

[1154] Step 7:

[1155] The server calculates the reference ratio. The input is the similarity score with existing images, and the output is the reference ratio. This process quantitatively calculates how much the uploaded image references existing images based on the similarity score.

[1156] Step 8:

[1157] The server analyzes the user's emotional data and adjusts the analysis results based on that emotional state. The input is the emotional data, and the output is the adjusted analysis results. In this process, the Emotion Engine is used to recognize the user's emotional state, and the analysis results reflect responses that match those emotions.

[1158] Step 9:

[1159] The server generates a report based on the judgment results, reference ratios, and emotion data and sends it to the user. The inputs are the judgment results, reference ratios, and emotion data, and the output is the generated report. In this process, the ReportGenerator aggregates the analysis results and creates and notifies the user of the report in an easy-to-understand format.

[1160] Through each of the above steps, this system is able to verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose appropriate ad placements based on user sentiment.

[1161] 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.

[1162] 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.

[1163] 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.

[1164] [Fourth embodiment]

[1165] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1166] 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.

[1167] 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).

[1168] 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.

[1169] 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.

[1170] 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).

[1171] 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.

[1172] 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.

[1173] 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.

[1174] 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.

[1175] 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.

[1176] 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.

[1177] 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."

[1178] 1. System Configuration

[1179] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[1180] 2. Program Overview

[1181] The server performs a series of analytical processes on the image data uploaded by the user, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, and analysis result report generation. The detailed operation of each step is explained below.

[1182] 3. User Operation

[1183] Users select the image data they wish to analyze and upload it to the server from their device. User operations are performed via an intuitive and simple interface.

[1184] 4. Server Processing Details

[1185] Pretreatment

[1186] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[1187] Feature Extraction

[1188] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[1189] AI generation judgment

[1190] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[1191] Identifying the source of the misappropriation

[1192] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[1193] Calculating the Reference Ratio

[1194] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[1195] 5. Generate and send the report

[1196] The server generates a detailed report based on the results of the judgment and the analysis of the reference ratio. This report includes data on the possibility of AI generation, the identification of the source of misuse, and the reference ratio. The generated report is sent from the server to the user, who can view it on their device.

[1197] Specific examples

[1198] For example, if a user uploads an image called "sunset.jpg," the server performs the following processes in order: Receive "sunset.jpg" and perform appropriate resizing and noise reduction. Next, extract feature vectors using edge detection and color histograms. After that, use a machine learning model to determine whether it is AI-generated and compare it with existing images in the database to identify the source. Finally, calculate the reference ratio between "sunset.jpg" and the most similar image, and send a report summarizing all these results to the user.

[1199] In this way, the present invention allows users to easily and efficiently obtain the origin and reference information of an image, making it easier to verify the authenticity of the image and ensure its originality. This system is highly useful in a variety of application fields.

[1200] The processing flow will be explained below.

[1201] Step 1: Upload user image

[1202] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data to the server.

[1203] Step 2: The server receives the image data

[1204] The server receives the image data sent by the user and checks the format of the image data to see if there are any problems.

[1205] Step 3: The server performs image preprocessing

[1206] The server performs preprocessing on the received image data, including image resizing and noise removal, to prepare the image for further analysis.

[1207] Step 4: The server extracts the feature vectors

[1208] The server extracts feature vectors from the preprocessed images, using techniques such as edge detection and color histograms to extract specific patterns and characteristics of the image as numerical data.

[1209] Step 5: The server determines the possibility of AI generation

[1210] The server inputs the extracted feature vector into a machine learning model, which evaluates the likelihood of AI generation and generates a probability score. The server then uses this probability score to determine whether the image was AI-generated.

[1211] Step 6: The server identifies potential sources of misappropriation

[1212] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[1213] Step 7: The server calculates the referral rate

[1214] The server calculates the reference ratio using the uploaded image and the candidate image for reuse. The reference ratio is a numerical value that indicates the degree to which a specific image references other images.

[1215] Step 8: Server Generates Report

[1216] The server generates a detailed report based on the AI-generated image judgment results and the analysis of the reference ratio, including the probability that the image was generated by AI, the detected potential sources of misappropriation, and the reference ratio.

[1217] Step 9: Server sends report to user

[1218] The server sends the generated report to the user, who receives the report on their device and checks the analysis results.

[1219] Example 1

[1220] 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."

[1221] Conventional image analysis systems have had difficulty efficiently determining whether image data uploaded by users is AI-generated, identifying the source of misuse, and calculating the reference ratio. In particular, it has been difficult to automatically and accurately perform a series of processes, from highly accurate preprocessing, feature extraction, AI generation determination, reference ratio calculation, and report generation. The purpose of this invention is to address these challenges and provide a system that allows users to easily verify the authenticity and originality of image data.

[1222] 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.

[1223] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This makes it possible to accurately and quickly determine the possibility of AI generation of image data uploaded by a user and identify the source of the misuse.

[1224] "Image data" refers to a digital file format that electronically stores visual information.

[1225] "Preprocessing" refers to initial processing such as resizing and noise removal performed on image data to improve the accuracy of analysis.

[1226] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[1227] "AI generation determination" refers to the process of using a machine learning model to determine whether an image was generated by artificial intelligence.

[1228] "Similarity" is a value calculated by calculating the degree of match between the feature vectors of two images.

[1229] The "reference ratio" is a quantitative ratio that indicates the extent to which an uploaded image references an existing image.

[1230] A "report" is a document summarizing the results of analysis such as the possibility of AI generation, identification of the source of misuse, and reference ratio.

[1231] A "database" is a collection of information that registers and manages a large number of known image data.

[1232] A "machine learning model" is an algorithm that is trained on a large amount of data in advance and then uses the results to analyze new data.

[1233] "Edge detection" is an image processing technique that identifies edges and boundaries within an image.

[1234] A "color histogram" is statistical data that represents the distribution of colors contained in an image.

[1235] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.

[1236] Hardware and software used

[1237] The following hardware and software are used to implement the present invention.

[1238] Hardware: Server machine equipped with high-speed CPU and GPU

[1239] Software: Image processing libraries (e.g., OpenCV), machine learning libraries (e.g., TensorFlow, PyTorch), database management systems (e.g., MySQL)

[1240] Example of operation

[1241] For example, if a user wants to upload an image called "sunset.jpg", the process is as follows:

[1242] 1. User operations

[1243] The user selects an image file "sunset.jpg" and uploads it to the server via the terminal. An example prompt is as follows:

[1244] Input image: sunset.jpg

[1245] 2. Server Reception

[1246] The server receives the uploaded "sunset.jpg" file and saves it in the specified directory. A message indicating that the file has been received is recorded in the server log.

[1247] 3. Pretreatment

[1248] The server performs preprocessing on the received image files, including image resizing and noise removal. Specifically, the server uses OpenCV to resize the image to a specific size and remove noise.

[1249] 4. Feature Extraction

[1250] After preprocessing, the server extracts feature vectors from the images using edge detection and color histograms, which represent specific patterns and characteristics of the image as numerical data.

[1251] 5. AI generation judgment

[1252] The extracted feature vectors are input into a machine learning model to determine whether the image is AI-generated or not, using a model trained in TensorFlow or PyTorch to obtain a probability score for the likelihood of AI generation.

[1253] 6. Identifying the source of the misappropriation

[1254] The server compares the feature vectors with existing images in a database, which contains a large number of known image data, and identifies the closest candidates based on the calculated similarity.

[1255] 7. Calculation of Reference Ratio

[1256] For each identified source image, the server calculates the reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[1257] 8. Generate and send reports

[1258] The server compiles information on the possibility of AI generation, the identification of the source of misuse, and the reference ratio to generate a detailed report. This report is generated in PDF or HTML format and sent to the user, who can then view the analysis results on their device.

[1259] The present invention provides a system that can accurately identify the possibility that image data uploaded by users is AI-generated and the source of the misuse, allowing users to easily verify the authenticity and originality of images.

[1260] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1261] Step 1: User selects and uploads image data

[1262] Users select the image data they wish to analyze from their own device. For example, they select an image file called "sunset.jpg" and upload it to the server via an intuitive interface. They select the "sunset.jpg" file as input and click the "Upload" button. The image data is sent to the server as output.

[1263] Step 2: The server receives the image data.

[1264] The server receives the "sunset.jpg" file uploaded by the user. As input, the server receives the image data sent from the user's device and saves it in the specified directory. Specifically, a message indicating completion of reception is recorded in the server's log. As output, the image data is saved on the server in a state where it can be preprocessed.

[1265] Step 3: Server preprocessing

[1266] The server performs preprocessing on the received image data. It reads the "sunset.jpg" file as input and performs the following operations: it resizes the image to a specific size using the OpenCV library (cv2.resize(image, (256, 256))) and removes noise (cv2.fastNlMeansDenoisingColored(image, None, 10, 10, 7, 21)). The output is the preprocessed image data.

[1267] Step 4: The server performs feature extraction

[1268] After preprocessing, the server extracts a feature vector from the image data. Using the preprocessed image data as input, it performs the following operations: edge detection (cv2.Canny(image, 100, 200)) and color histogram (cv2.calcHist([image], [0, 1, 2], None, [8, 8, 8], [0, 256, 0, 256, 0, 256])). The output is the feature vector of the image.

[1269] Step 5: The server makes the AI ​​generation decision

[1270] The server inputs the extracted feature vector into a machine learning model to determine whether the image was generated by AI. The feature vector is input into the machine learning model as input, and the following process is performed: Specifically, the feature vector is analyzed using a pre-trained model (e.g., TensorFlow or PyTorch) (model(features)). The output is a probability score for the likelihood of AI generation.

[1271] Step 6: The server identifies the source of the stolen data

[1272] The server compares the extracted feature vector with existing images in the database. As input, it uses the feature vector to calculate the similarity with existing images in the database. Specifically, it performs the similarity calculation and identifies the candidate source with the closest match (compare_with_database(features)). As output, it identifies the most similar existing image.

[1273] Step 7: The server calculates the referral rate.

[1274] Based on the identified source image, the server calculates the reference ratio. Using the feature vectors of the identified source image and the uploaded image data as input, the server executes the following process. Specifically, it calculates the reference ratio with the source image (calculate_reference_ratio(features, best_match_features)). The output shows the reference ratio quantitatively.

[1275] Step 8: The server generates the report and sends it to the user

[1276] The server generates a detailed report based on the analysis results and sends it to the user. It uses the AI-generated probability score, misappropriation source identification, and reference ratio data as input and performs the following process: Specifically, it generates a report in PDF or HTML format (generate_report(probability_score, best_match_details, reference_ratio)) and sends it to the user (send_report_to_user(report)). The analysis results are provided to the user as output.

[1277] (Application example 1)

[1278] 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."

[1279] In modern society, the ease with which counterfeit and inappropriate images are spread over the Internet is increasing, and these false information and inappropriate images can cause serious problems for individuals, companies, and organizations. Therefore, there is a need to quickly and accurately determine the authenticity of images and take appropriate countermeasures as necessary. However, with current technology, determining the authenticity of images and detecting counterfeits often requires time and effort, making real-time detection difficult. This presents a challenge in terms of making it difficult to respond quickly and effectively.

[1280] 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.

[1281] In this invention, the server includes means for receiving and preprocessing image data, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for determining whether the image data is forged, and means for generating a report based on the determination result and the reference ratio and sending it to the user. This enables users to quickly and accurately determine whether images on the Internet are forged and take appropriate measures in real time based on the results.

[1282] "Image data" is visual information represented in digital form.

[1283] "Preprocessing" refers to preparatory work to improve the accuracy of analysis, such as resizing image data and removing noise.

[1284] A "feature vector" is a numerical representation of a particular pattern or characteristic of an image.

[1285] "Likelihood of AI Generation" is a probability score that indicates whether an image was generated by artificial intelligence.

[1286] The "similarity" is a value calculated by using feature vectors to determine the similarity between different images.

[1287] The "reference ratio" is a quantitative index that indicates the degree to which an image references an existing image.

[1288] "Counterfeit determination" is the process of assessing whether an image is counterfeit.

[1289] A "report" is a document summarizing the analysis results and is information provided to the user.

[1290] A "server" is a computer system that processes data on a network.

[1291] A "user" is an individual or organization that uses the system to upload image data and receive analysis results.

[1292] The present invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio. This system is particularly useful for detecting counterfeit images in security services.

[1293] 1. System Configuration

[1294] The present invention mainly comprises the following elements:

[1295] Hardware

[1296] Device: A device that takes and uploads images, such as a smartphone or smart glasses.

[1297] Server: A computer system that receives image data and performs analysis and processing.

[1298] software

[1299] Python: A programming language for writing the overall program logic.

[1300] TensorFlow: A library for extracting feature vectors and running machine learning models to determine AI-generated possibilities.

[1301] Sklearn: A library for computing similarities between feature vectors.

[1302] 2. Specific processing flow

[1303] User operations

[1304] Users use the terminal to select the image data they wish to analyze and upload it to the server. The terminal has an intuitive and simple interface, so users can operate it without any special skills.

[1305] Server Processing

[1306] 1. Pretreatment

[1307] The server receives the image data sent by the user and performs preprocessing such as resizing and noise removal on the received image data.

[1308] 2. Extracting feature vectors

[1309] After preprocessing is complete, feature vectors are extracted from the image data using techniques such as edge detection and color histograms.

[1310] 3. AI-generated judgment

[1311] The extracted feature vector is input into a machine learning model to determine whether the image was generated by AI. The machine learning model is built using TensorFlow and is pre-trained with a large amount of data.

[1312] 4. Identifying the source of the misappropriation

[1313] The feature vector is matched with existing images in a database to identify the most similar images. Similarity is calculated using the cosine_similarity function in Sklearn.

[1314] 5. Calculation of reference ratio

[1315] The reference ratio is calculated based on the similarity between the original image and the uploaded image.

[1316] 6. Report Generation and Notifications

[1317] A detailed report is generated based on the assessment results and reference ratio and sent to the user, who can then check the report on their device and take any necessary measures based on the information.

[1318] Specific examples

[1319] If a user takes an image titled "Important Document.png" during an important company meeting and wants to check whether the image may be forged, they can use this system. The user uploads the image to the server, which performs a series of analysis processes and then provides the user with a determination of whether the image is forged.

[1320] Prompt Sentence Examples

[1321] "Determine if this image was generated by AI, identify which existing images it references, and generate a report on the likelihood that the image is truly original."

[1322] This system allows users to quickly and accurately determine the authenticity of images and take necessary measures in real time.

[1323] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1324] Step 1:

[1325] The user uses the device to select the image data they wish to analyze and upload it to the server. The input is the image data selected by the user operating the device, and the output is the image data sent to the server. Specifically, the user selects "Important Document.png" from the smartphone app and presses the upload button.

[1326] Step 2:

[1327] The server receives the image data sent by the user. The input is the image data sent in step 1, and the output is the image data saved in the server. Specifically, the server receives the HTTP request and saves the image file in a save directory.

[1328] Step 3:

[1329] Preprocessing is performed. The server resizes the received image data and removes noise. The input is the stored image data, and the output is the preprocessed image data. Specifically, the image is resized to 256x256 pixels and noise filtering is applied.

[1330] Step 4:

[1331] Extract feature vectors. The server generates feature vectors from the preprocessed image data using edge detection and color histograms. The input is the preprocessed image data, and the output is a feature vector. Specifically, it extracts edges from the image and quantifies the color distribution using a color histogram.

[1332] Step 5:

[1333] Determine the possibility of AI generation. The server inputs the feature vector into a machine learning model to determine whether the image was generated by AI. The input is the feature vector, and the output is a probability score of AI generation. Specifically, the feature vector is input into the TensorFlow model and the determination result is obtained.

[1334] Step 6:

[1335] The source of the image is identified. The server compares the feature vector with images in an existing database and identifies the most similar image. The input is the feature vector, and the output is the most similar existing image and its similarity. Specifically, the similarity is calculated using Sklearn's cosine_similarity function, and the index of the maximum value is obtained.

[1336] Step 7:

[1337] Calculate the reference ratio. The server calculates the reference ratio based on the similarity with the identified source image. The input is the similarity with the source image, and the output is the reference ratio. Specifically, the server expresses the reference ratio as a percentage based on the similarity score.

[1338] Step 8:

[1339] A report is generated and sent to the user. The server creates a detailed report based on the judgment results and reference ratio and sends it to the user's device. The input is the AI-generated likelihood score and reference ratio, and the output is report data. Specifically, the analysis results are summarized in text format and sent to the user via email or app notification.

[1340] This allows users to immediately check the analysis results and determine the authenticity of the image.

[1341] 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.

[1342] 1. System Configuration

[1343] This invention is a system in which a user uploads image data to a server using a terminal, and the server analyzes the image data to determine the possibility of AI generation, identify the reference source, and analyze the reference ratio.It also combines an emotion engine that recognizes the user's emotions.

[1344] 2. Program Overview

[1345] The server performs a series of analytical processes on image data uploaded by users, including preprocessing, feature extraction, AI generation determination, source identification, reference ratio calculation, analysis result report generation, and user emotion recognition using an emotion engine. The detailed operation of each step is explained below.

[1346] 3. User Operation

[1347] The user selects the image data they wish to analyze and uploads it to the server from their device. The device also captures the user's facial expressions and voice data for emotion recognition. User operation is performed via an intuitive and simple interface.

[1348] 4. Server Processing Details

[1349] Pretreatment

[1350] The server receives the image data sent by the user and performs preprocessing on the received image data, which includes image resizing and noise removal, and is an important step to improve the accuracy of analysis.

[1351] Feature Extraction

[1352] After preprocessing is complete, the server extracts feature vectors from the image data. Feature extraction uses techniques such as edge detection and color histograms, allowing specific patterns and characteristics of the image to be expressed as numerical data.

[1353] AI generation judgment

[1354] The server inputs the extracted feature vectors into a machine learning model to evaluate the likelihood of AI generation. The machine learning model is pre-trained with a large amount of data and can make highly accurate judgments. The result is a probability score for whether the image was generated by AI.

[1355] Identifying the source of the misappropriation

[1356] The server then matches the feature vector with existing images in a database. Based on the calculated similarity, the server identifies the closest match. This database contains a large number of known image data, allowing for effective comparison.

[1357] Calculating the Reference Ratio

[1358] For each identified source image, the server calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references an existing image.

[1359] 5. Emotion Engine Processing

[1360] Capturing Emotional Data

[1361] The device captures the user's facial and voice data and sends it to the emotion engine, which then analyzes the user's facial and voice data to recognize their emotional state.

[1362] Emotion analysis

[1363] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, and this emotion data is reflected in the analysis results.

[1364] 6. Generate and send reports

[1365] The server generates a detailed report based on the judgment results, reference ratio, and the emotion engine's recognition results. The report includes the possibility that the image was generated by AI, the detected source of misappropriation, the reference ratio, and a corresponding message based on the user's emotion. The report output format is adjusted taking this data into account.

[1366] Specific examples

[1367] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following process: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms. Then, use a machine learning model to determine whether the image is AI-generated and match it with existing images in the database to identify the source. Calculate the reference ratio and generate a detailed report. Finally, the report output format is adjusted based on the user's emotional state of excitement, and sent from the server to the user.

[1368] In this way, by combining the emotion engine, flexible report creation and notification based on the user's emotions becomes possible, making the way analysis results are received more user-friendly. This system is highly useful in a variety of fields.

[1369] The processing flow will be explained below.

[1370] Step 1: Upload user image

[1371] The user selects the image data they want to analyze using their device. When the user clicks the "Upload" button, the device sends the image data along with the user's facial expression and voice data to the server.

[1372] Step 2: The server receives the image data

[1373] The server receives the image data sent by the user, checks the format of the image data, and determines whether it can be analyzed.

[1374] Step 3: The server performs image preprocessing

[1375] The server performs preprocessing on the received image data. This includes image resizing and noise removal. Preprocessing is a necessary preparation to improve the accuracy of analysis.

[1376] Step 4: The server extracts the feature vectors

[1377] The server uses techniques such as edge detection and color histograms to extract feature vectors from the pre-processed images, which represent specific patterns or characteristics of the image as numerical data.

[1378] Step 5: The server determines the possibility of AI generation

[1379] The server uses the extracted feature vector to input data into a machine learning model, which evaluates the likelihood of AI generation and calculates a probability score. Based on this probability score, the server determines whether the image is AI-generated.

[1380] Step 6: The server identifies potential sources of misappropriation

[1381] The server compares the feature vector with existing images in the database, performs a similarity calculation, and identifies the most similar image. This identified image becomes a candidate for reuse.

[1382] Step 7: The server calculates the referral rate

[1383] The server compares the uploaded image with the potential source image and calculates the reference ratio, which clarifies the extent to which the uploaded image references other images.

[1384] Step 8: The device captures emotion data

[1385] The device captures the user's facial expressions and voice data and sends it to the emotion engine, which allows the user's emotional state to be analyzed in real time.

[1386] Step 9: The server analyzes the emotion data

[1387] The server uses an emotion engine to analyze facial expressions and voice data sent from the device, and classifies the user's emotions into categories such as "happiness," "surprise," "sadness," and "anger."

[1388] Step 10: Server Generates Report

[1389] The server generates a detailed report based on the AI-generated judgment results, reference ratio, and the emotion engine's recognition results, including the possibility that the image was generated by AI, the detected potential sources of misappropriation, the reference ratio, and a corresponding message according to the user's emotion.

[1390] Step 11: Server sends report to user

[1391] The server sends the generated report to the user, who receives the report on their device and checks the analysis results and corresponding messages.

[1392] Example 2

[1393] 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."

[1394] Conventional image analysis systems can perform preprocessing of image data, extraction of feature vectors, determination of the possibility of AI generation, calculation of similarity, and calculation of reference ratios, but they lack the functionality to reflect the user's emotional state in the analysis results, which means that the user experience is not sufficiently improved.Another problem is that the reports generated based on the analysis results are uniform, making it impossible to respond flexibly to the user's emotions.

[1395] 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.

[1396] In this invention, the server includes means for receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for generating a report based on the determination result and the reference ratio and sending it to the user, means for receiving the user's facial expression and voice data and recognizing their emotional state, and means for adjusting the report output format based on the recognized emotional state. This makes it possible to generate reports that reflect the user's emotional state, making it more user-friendly to receive analysis results.

[1397] "Image data" means visual information represented in digital form.

[1398] "Preprocessing" refers to initial data processing operations performed to improve the accuracy of the analysis of image data.

[1399] A "feature vector" is a numerical representation of characteristics and patterns extracted from image data.

[1400] "AI-generated likelihood determination" is the process of using the extracted feature vectors to determine whether an image was generated by artificial intelligence.

[1401] "Calculation of similarity" is a process of numerically evaluating the similarity between image data and existing image data based on the feature vector.

[1402] The "reference ratio" is a ratio that quantitatively indicates to what extent the uploaded image references existing image data.

[1403] "Report generation" is the process of creating a detailed report based on the analysis results to provide to the user.

[1404] "User's facial expression and voice data" refers to data that digitally represents the user's facial expression and vocal content.

[1405] "Emotional state recognition" is the process of analyzing a user's facial expressions and voice data to determine the user's emotional state.

[1406] "Adjusting report output format" refers to flexibly changing the appearance and content of generated reports based on the perceived emotional state.

[1407] In this system, users upload image data to a server using their devices, and the server analyzes the image data to identify the possibility of AI generation, the reference source, and the reference ratio. Furthermore, it combines an emotion engine that recognizes the user's emotions.

[1408] 1. System Configuration

[1409] The system consists of the following main components:

[1410] Terminal

[1411] server

[1412] Emotion Engine

[1413] 2. User Operation

[1414] Users select the image data they wish to analyze and upload it to the server from their device. In addition to the image data, the device also captures the user's facial expressions and voice data and sends this data to the server. Users operate the system via an intuitive and simple interface.

[1415] 3. Image data processing details (server operation)

[1416] The server processes the received image data in the following steps:

[1417] 1. Pretreatment

[1418] The server resizes the images and removes noise, specifically using libraries such as OpenCV.

[1419] 2. Extracting feature vectors

[1420] Extract feature vectors from preprocessed image data using edge detection (e.g., Canny method) and color histograms. Use functions such as cv2.calcHist.

[1421] 3. AI-generated possibility assessment

[1422] The feature vector is fed into a machine learning model (e.g., a TensorFlow model) to assess the likelihood that the image was generated by AI.

[1423] 4. Identifying the source of the misappropriation

[1424] The feature vectors are matched with existing images in a database, and potential sources are identified based on similarity (e.g., cosine similarity).

[1425] 5. Calculation of Reference Ratio

[1426] The reference ratio is calculated for the original image, and the extent to which the existing image is referenced is quantitatively shown.

[1427] 4. Emotion engine processing (terminal and server operation)

[1428] 1. Capturing Emotional Data

[1429] The device captures the user's facial expressions and voice data and sends that data to the emotion engine.

[1430] 2. Emotion analysis

[1431] The server's emotion engine analyzes the received facial and voice data to recognize the user's emotions in real time, using facial recognition algorithms and voice emotion recognition models.

[1432] 5. Generate and send reports (server operations)

[1433] The server generates a detailed report based on the judgment results, reference ratios, and the emotion engine's recognition results, including:

[1434] The image may have been generated by AI

[1435] Identified potential sources of misappropriation

[1436] Reference ratio

[1437] Response messages based on user sentiment

[1438] This report is sent to the user, and the output format is adjusted according to the user's feelings.

[1439] Specific examples

[1440] For example, if a user uploads an image called "sunset.jpg" and the emotion engine detects "excitement," the server performs the following processes: Receive "sunset.jpg" and resize and remove noise. Next, extract feature vectors using edge detection and color histograms, and use a machine learning model to determine the possibility of AI generation. Match the image with existing images in the database to identify the source and calculate the reference ratio. Generate a detailed report, adjust the report output format based on the user's emotional state of excitement, and send it to the user from the server.

[1441] Example prompts for generative AI models

[1442] "Please determine whether the uploaded image data was generated by AI and analyze whether it has a specific source. Also, please determine the user's emotion and generate a report based on that. For example, the image file name is 'sunset.jpg' and the user's emotion is 'excited'."

[1443] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1444] Step 1: Upload image data

[1445] Input: An image file selected by the user on the device.

[1446] Processing: The user selects the image data they wish to analyze and uploads it from their device to the server.

[1447] Output: Image file uploaded to the server

[1448] Specific operation: The user selects the desired image file (e.g., "sunset.jpg") from the device's file system and clicks the "Upload" button, which sends the image data to the server.

[1449] Step 2: Capturing emotion data

[1450] Input: User facial and voice data

[1451] Processing: The device captures the user's facial expressions and voice and sends them to the server as emotional data.

[1452] Output: Facial expression and voice data sent to the server

[1453] Specific operation: The device's camera and microphone are used to record and photograph the user's face and voice in real time, and the recorded and photographed data is sent to a server via the network.

[1454] Step 3: Preprocessing the image data

[1455] Input: Image file uploaded to the server

[1456] Processing: The server performs preprocessing on the received image data.

[1457] Output: Preprocessed image data

[1458] Specific operation: The server uses a library (e.g. OpenCV) to resize the image data to a standard size and remove noise using filtering techniques.

[1459] Step 4: Extract feature vectors

[1460] Input: Preprocessed image data

[1461] Processing: The server extracts feature vectors from the preprocessed image data.

[1462] Output: Extracted feature vector

[1463] Specific operation: The server uses edge detection (e.g., Canny method) and color histogram (e.g., cv2.calcHist function) to extract image features as numerical data.

[1464] Step 5: Determine the possibility of AI generation

[1465] Input: Extracted feature vector

[1466] Processing: The server inputs the extracted feature vectors into a machine learning model to determine the likelihood that the image was generated by AI.

[1467] Output: Probability score indicating the likelihood of the AI ​​generation

[1468] What it does: The server uses a trained machine learning model (e.g., a TensorFlow model) to evaluate feature vectors and output an AI-generated probability score.

[1469] Step 6: Identify the source

[1470] Input: Extracted feature vector

[1471] Processing: The server matches the feature vector with existing images in a database to identify potential sources.

[1472] Output: Identified potential sources of misappropriation

[1473] Specific operation: The server queries the database for the feature vector and identifies the closest match based on a calculated similarity (e.g., cosine similarity).

[1474] Step 7: Calculate the reference ratio

[1475] Input: Uploaded image file, identified potential source

[1476] Processing: The server calculates the reference ratio for the identified source image.

[1477] Output: Reference ratio

[1478] Specific operation: The server calculates the reference ratio based on the similarity between the uploaded image and the original image, and quantitatively indicates the extent to which the image has been referenced.

[1479] Step 8: Analyze the sentiment data

[1480] Input: Facial and voice data sent to the server

[1481] Processing: The server's emotion engine analyzes the received facial and voice data to recognize the user's emotional state.

[1482] Output: User's emotional state

[1483] Specific operation: The server uses an emotion analysis algorithm (e.g., a facial recognition algorithm or a voice emotion recognition model) to analyze emotions from the user's facial expressions and voice.

[1484] Step 9: Generate reports

[1485] Input: Probability score indicating the likelihood of AI generation, identified potential sources of misappropriation, reference ratio, and user emotional state

[1486] Processing: The server generates a report based on the judgment results, reference ratio, and the emotion engine's recognition results.

[1487] Output: Detailed report for the user

[1488] Specific operation: The server organizes the analysis results and outputs a report in a format that is easy for the user to understand (e.g., PDF or HTML format). The report includes the possibility of the image being generated by AI, potential sources of reuse, the reference ratio, and a corresponding message based on the user's emotions.

[1489] Step 10: Submit the report

[1490] Input: Detailed reports for users

[1491] Processing: The server generates the report and sends it to the user.

[1492] Output: Report sent to user

[1493] What happens: The server sends the report to the user via email or a download link. The user can download the report by clicking the provided link.

[1494] (Application example 2)

[1495] 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."

[1496] There is a need to verify the originality of image materials used by advertising creators and optimize ad placement based on user emotions, but conventional systems have had difficulty doing this effectively.In addition, there was no system that could flexibly respond by taking user emotions into account when evaluating the possibility that an image was generated by AI or identifying the source of the image.

[1497] 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 receiving image data and performing preprocessing, means for extracting feature vectors from the preprocessed image data, means for determining the possibility of AI generation using the extracted feature vectors, means for comparing the feature vectors with existing images in a database and calculating similarity, means for calculating a reference ratio, means for capturing user emotion data and adjusting the analysis results based on the emotion data, and means for generating a report based on the determination results, reference ratio, and emotion data and sending it to the user. This makes it possible to confirm the originality of image materials used by advertising creators, evaluate the possibility of AI generation, and identify sources of reuse, as well as to suggest optimal ad placement based on user emotion.

[1498] "Image data" is digital data of visual information that a user uploads to a server using a terminal.

[1499] "Preprocessing" refers to initial processing such as resizing and noise removal that is performed before analyzing image data.

[1500] A "feature vector" is a numerical representation of a particular pattern or characteristic of image data.

[1501] "Likelihood of AI Generation" is a probability score that assesses whether image data was generated by artificial intelligence.

[1502] "Similarity" is a numerical representation of the degree of match between feature vectors.

[1503] The "reference ratio" is the percentage that indicates the extent to which the uploaded image references an existing image.

[1504] "Emotion data" is information indicating the emotional state of the user analyzed from facial expressions and voice data.

[1505] "Capture" refers to the device recording the user's facial expressions and voice and acquiring them as digital data.

[1506] The "analysis results" are a compilation of all the processing results from the server's preprocessing of image data, feature extraction, AI generation judgment, similarity calculation, reference ratio calculation, and emotion analysis.

[1507] A "report" is a detailed report generated based on the analysis results and emotion data and provided to the user.

[1508] 1. System Configuration

[1509] This invention is a system aimed at verifying the originality of image materials used by advertising creators and optimizing advertising placement based on user emotions. The system consists of three main elements: a server, a terminal, and a user. The server analyzes image data, the terminal captures image data and emotional data from the user, and the user uploads the image data.

[1510] 2. Program Overview

[1511] To implement this invention, the following hardware and software are used. The hardware requires a terminal equipped with a camera for capturing images and a microphone for capturing audio, and a server capable of high-speed computation. The software requires a library using OpenCV (cv2) to resize and denoise images, an engine using EmotionEngine to recognize user emotions, a module using FeatureExtractor to extract feature vectors, a module using AIDetector to determine the possibility of AI generation, a module using SimilaritySearch to identify the source of reuse, and a module using ReportGenerator to generate the final report.

[1512] 3. Processing Flow

[1513] The user selects the image data they wish to analyze and uploads it from their device to the server. The device simultaneously captures the user's facial expressions and voice data and transmits their emotional state to the server. The server first performs preprocessing on the received image data, including image resizing and noise removal. After the preprocessing is complete, feature vectors are extracted from the image data. Edge detection and color histograms are used to extract the feature vectors.

[1514] The server then uses the extracted feature vectors to assess the likelihood of AI generation. This is done using a machine learning model previously trained on a large amount of data. The result of the AI ​​generation assessment is a probability score indicating whether the image was generated by AI. The server then matches the feature vectors with existing images in its database to identify the source of the image. For each identified source image, it calculates a reference ratio, which quantitatively indicates the extent to which the uploaded image references existing images.

[1515] Furthermore, the emotion engine analyzes the user's emotional state and reflects this emotional data in the analysis results. Finally, the server generates a detailed report based on the analysis results, reference ratio, and emotion data and sends it to the user. The report includes the possibility that the image was generated by AI, the original image and reference ratio, and a corresponding message based on the user's emotion.

[1516] 4. Specific Examples

[1517] For example, consider the case where an ad creator uploads a new ad banner image "ad_banner.jpg" to the app and simultaneously captures user sentiment data (interestingness). The server receives this image, resizes it, and denoises it. Next, edge detection and color histograms are used to extract feature vectors. A machine learning model is then used to determine whether it is AI-generated and match it with existing images in the database to identify the source. The reference ratio is calculated and a detailed report is generated. Finally, a report containing optimal ad placement suggestions based on user sentiment is generated and sent to the creator.

[1518] Example prompt sentence:

[1519] "Upload your ad banner image and see how original and AI-generated it is."

[1520] "Please suggest the best ad placement if the user finds it interesting."

[1521] In this way, this system can verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose optimal ad placements based on user sentiment.

[1522] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1523] Step 1:

[1524] The user selects the image data they wish to analyze and uploads it from their device to the server. The input is the image data "ad_banner.jpg" selected by the user, and the output is the image data sent to the server. In this process, the device reads the image data and sends it to the specified server's upload endpoint.

[1525] Step 2:

[1526] The device captures the user's facial and voice data and sends the emotion data to the server. The input is the user's facial and voice data, and the output is the emotion data sent to the server. In this process, the device captures data in real time using a camera and microphone and sends it to the server in a format for emotion analysis.

[1527] Step 3:

[1528] The server performs preprocessing on the received image data. The input is the image data "ad_banner.jpg" sent to the server, and the output is the image data after resizing and noise removal. In this process, OpenCV (cv2) is used to resize the image to 256x256 pixels and apply a noise removal algorithm.

[1529] Step 4:

[1530] The server extracts feature vectors from the preprocessed image data. The input is the preprocessed image data, and the output is a feature vector. In this process, FeatureExtractor is used to apply edge detection and color histogram techniques to extract image characteristics as numerical data.

[1531] Step 5:

[1532] The server uses the extracted feature vector to determine the possibility of AI generation. The input is the extracted feature vector, and the output is a probability score indicating the possibility of AI generation. In this process, the feature vector is input into the pre-trained AIDetector machine learning model to evaluate the possibility of AI generation.

[1533] Step 6:

[1534] The server compares the feature vector with existing images in the database and calculates the similarity. The input is the extracted feature vector, and the output is the similarity score with the existing image and the candidate image to be reused. In this process, SimilaritySearch is used to compare the feature vector with images in the database and identify the most similar image.

[1535] Step 7:

[1536] The server calculates the reference ratio. The input is the similarity score with existing images, and the output is the reference ratio. This process quantitatively calculates how much the uploaded image references existing images based on the similarity score.

[1537] Step 8:

[1538] The server analyzes the user's emotional data and adjusts the analysis results based on that emotional state. The input is the emotional data, and the output is the adjusted analysis results. In this process, the Emotion Engine is used to recognize the user's emotional state, and the analysis results reflect responses that match those emotions.

[1539] Step 9:

[1540] The server generates a report based on the judgment results, reference ratios, and emotion data and sends it to the user. The inputs are the judgment results, reference ratios, and emotion data, and the output is the generated report. In this process, the ReportGenerator aggregates the analysis results and creates and notifies the user of the report in an easy-to-understand format.

[1541] Through each of the above steps, this system is able to verify the originality of image materials used by advertising creators, evaluate the possibility of AI generation, identify the source of misuse, and propose appropriate ad placements based on user sentiment.

[1542] 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.

[1543] 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.

[1544] 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.

[1545] 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.

[1546] 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.

[1547] 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.

[1548] 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).

[1549] 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.

[1550] 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."

[1551] 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.

[1552] 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).

[1553] 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.

[1554] 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.

[1555] 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.

[1556] 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.

[1557] 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.

[1558] 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.

[1559] 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.

[1560] 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.

[1561] 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.

[1562] 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.

[1563] The following is further disclosed regarding the above embodiment.

[1564] (Claim 1)

[1565] means for receiving and pre-processing image data;

[1566] means for extracting feature vectors from the preprocessed image data;

[1567] A means for determining the possibility of generating AI using the extracted feature vector;

[1568] means for matching the feature vector with existing images in a database and calculating the similarity;

[1569] a means for calculating a reference ratio;

[1570] means for generating a report based on the determination result and the reference ratio and sending the report to a user;

[1571] A system including:

[1572] (Claim 2)

[1573] 10. The system of claim 1, wherein the preprocessing includes image resizing and noise removal.

[1574] (Claim 3)

[1575] 2. The system of claim 1, wherein the feature vector is extracted using edge detection and color histograms.

[1576] "Example 1"

[1577] (Claim 1)

[1578] means for receiving and pre-processing image data;

[1579] means for extracting feature vectors from the preprocessed image data;

[1580] A means for determining the possibility of generating AI using the extracted feature vector;

[1581] means for matching the feature vector with existing images in a database and calculating the similarity;

[1582] a means for calculating a reference ratio;

[1583] means for generating a report based on the determination result and the reference ratio and sending the report to a user;

[1584] A system including:

[1585] (Claim 2)

[1586] 10. The system of claim 1, wherein the preprocessing includes image resizing and noise removal.

[1587] (Claim 3)

[1588] 2. The system of claim 1, wherein the feature vector is extracted using edge detection and color histograms.

[1589] (Claim 4)

[1590] A means for a user to upload image data that the user wishes to analyze to the system via a terminal;

[1591] A means for the server to receive the uploaded image data;

[1592] The system of claim 1 further comprising:

[1593] (Claim 5)

[1594] A means for inputting the feature vector extracted by the server into a machine learning model based on the preprocessed image data to obtain a probability score for the possibility of AI generation;

[1595] Machine learning models are pre-trained on a large number of datasets,

[1596] The system of claim 1 further comprising:

[1597] (Claim 6)

[1598] a means for the server to match the feature vectors with existing images in the database and identify the candidate appropriation source with the closest match;

[1599] A means for calculating a reference ratio for the identified reused original image;

[1600] The system of claim 1 further comprising:

[1601] (Claim 7)

[1602] A means for the server to generate a detailed report based on the analysis results and send it to the user;

[1603] The report will include data on the possibility of AI generation, identification of the source of misappropriation, and reference ratios.

[1604] The system of claim 1 further comprising:

[1605] "Application Example 1"

[1606] (Claim 1)

[1607] means for receiving and pre-processing image data;

[1608] means for extracting feature vectors from the preprocessed image data;

[1609] A means for determining the possibility of generating AI using the extracted feature vector;

[1610] means for matching the feature vector with existing images in a database and calculating the similarity;

[1611] a means for calculating a reference ratio;

[1612] means for determining whether the image data is forged;

[1613] means for generating a report based on the determination result and the reference ratio and sending the report to a user;

[1614] A system including:

[1615] (Claim 2)

[1616] 10. The system of claim 1, wherein the preprocessing includes image resizing and noise removal.

[1617] (Claim 3)

[1618] 2. The system of claim 1, wherein the feature vector is extracted using edge detection and color histograms.

[1619] "Example 2: Combining Emotion Engines"

[1620] (Claim 1)

[1621] means for receiving and pre-processing image data;

[1622] means for extracting feature vectors from the preprocessed image data;

[1623] A means for determining the possibility of generating AI using the extracted feature vector;

[1624] means for matching the feature vector with existing images in a database and calculating the similarity;

[1625] a means for calculating a reference ratio;

[1626] means for generating a report based on the determination result and the reference ratio and sending the report to a user;

[1627] means for receiving facial and voice data of a user and recognizing an emotional state;

[1628] means for adjusting the report output format based on the recognized emotional state;

[1629] A system including:

[1630] (Claim 2)

[1631] 10. The system of claim 1, wherein the preprocessing includes image resizing and noise removal.

[1632] (Claim 3)

[1633] 2. The system of claim 1, wherein the feature vector is extracted using edge detection and color histograms.

[1634] "Application example 2 when combining emotion engines"

[1635] (Claim 1)

[1636] means for receiving and pre-processing image data;

[1637] means for extracting feature vectors from the preprocessed image data;

[1638] A means for determining the possibility of generating AI using the extracted feature vector;

[1639] means for matching the feature vector with existing images in a database and calculating the similarity;

[1640] a means for calculating a reference ratio;

[1641] means for capturing user emotion data and adjusting analysis results based on the emotion data;

[1642] means for generating a report based on the judgment result, the reference ratio, and the emotion data and sending the report to the user;

[1643] A system including:

[1644] (Claim 2)

[1645] 10. The system of claim 1, wherein the preprocessing includes image resizing and noise removal.

[1646] (Claim 3)

[1647] 2. The system of claim 1, wherein the feature vector is extracted using edge detection and color histograms. [Explanation of symbols]

[1648] 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. means for receiving and pre-processing image data; means for extracting feature vectors from the preprocessed image data; A means for determining the possibility of generating AI using the extracted feature vector; means for matching the feature vector with existing images in a database and calculating the similarity; a means for calculating a reference ratio; means for generating a report based on the determination result and the reference ratio and sending the report to a user; A system including:

2. The system of claim 1 , wherein the preprocessing includes image resizing and denoising.

3. The system of claim 1 , wherein the extraction of the feature vector uses edge detection and a color histogram.

Images