An AI-based image evaluation and screening method and system

Abstract

The present invention discloses an AI-based image evaluation and screening method and system thereof, comprising acquiring image data in an album and grouping them into target image data of different groups, and generating a corresponding target image data set; calling the target image data set through an AI big model, performing AI intelligent analysis on the image screening influencing feature factor set based on the AI ​​big model, and generating image evaluation result data; constructing a screening model for images in the album to analyze and obtain target screening image result data; and performing corresponding screening on the image data in the album. The present invention applies AI to screen image quality, effectively realizing a more prepared judgment and evaluation of images after grading, which is conducive to accurately screening out poor images, and can intelligently screen out more accurate, beautiful and visually appealing images, further improving the hierarchical judgment effect of image classification.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to an AI-based picture evaluation and screening method and system thereof. Background Art

[0002] After traveling, people often filter the photos they took during the trip and select the best photos of the same scene to collect or share. At this time, they need to view the photos one by one, which takes a long time. Long-term screening will cause visual fatigue for users, making it difficult to filter out the best photos. Currently, many AI models have been applied to processing such as taking photos, such as screening photos for similarity processing. Currently, there are few applications that use AI technology to grade and screen the quality of pictures, especially the poor conditions of people and environments;

[0003] For example, in the publication number CN118154571A, which is named as an image quality evaluation method, system, device and medium for AI images, it includes: obtaining an initial AI image and corresponding text prompt data; preprocessing the initial AI image to obtain several AI images of different scales; inputting the AI ​​images of different scales and the corresponding text prompt data into a multi-scale image quality evaluation model for quality evaluation to obtain an image quality evaluation score; wherein the image quality evaluation score includes the content consistency score, visual quality score and authenticity score of the AI ​​image; the multi-scale image quality evaluation model includes a feature extraction module, an AFF module and an evaluation module connected in sequence; the feature extraction module is constructed based on an image encoder, and the evaluation module includes several fully connected layers arranged in parallel. The technical solution described in the present invention shows strong performance in the multi-dimensional evaluation of AI-generated images.

[0004] Publication No. CN118115831A, entitled "Intelligent Processing Method of Image Data Based on AI Training Platform", involves image intelligent technology, and proposes this solution to address the problem in the prior art that the volume of sample data affects the quality of the model. First, the collected sample data is screened through a data screening module based on the AI ​​training platform, and then the screened data is labeled; then the sample data is preprocessed, and during the preprocessing process, the sample data is enhanced using a data enhancement module to improve the generalization ability. The advantage is that it is done from two aspects: one is to select a large amount of data through a preset model; the other is to use a data enhancement module to enhance small sample data, further improve the generalization ability of the model, and increase the sample size. This method can save labor input and improve generalization ability.

[0005] In the face of the current image processing, accurate grading, evaluation and intelligent screening of image quality are required. For this purpose, we propose an AI-based image evaluation and screening method and system. Summary of the invention

[0006] In view of the above problems existing in the existing image processing, the present invention is proposed.

[0007] Therefore, one of the objects of the present invention is to provide an AI-based image evaluation and screening method and system thereof, which utilizes AI to perform image quality screening, can intelligently screen out more accurate, beautiful and visually appealing images, and combines hesitant fuzzy evaluation to improve the image grading evaluation method, thereby further improving the hierarchical evaluation effect of image classification.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0009] In a first aspect, the present invention provides an AI-based image evaluation and screening method and system thereof, comprising:

[0010] Acquire image data in the photo album, perform feature extraction, cluster analysis and scene time recognition on the image data in sequence based on classification features of the image data, group the image data into target image data of different groups according to the scene time classification features, and generate corresponding target image data sets;

[0011] The target image data set is called by the AI ​​big model, and after data classification, training, recognition and association are performed on the input target image data set, the image screening influencing features and the corresponding image screening influencing feature factor set in the image screening process are determined, and the image screening influencing feature factor set is subjected to AI intelligent analysis based on the AI ​​big model to generate result data of image evaluation;

[0012] Obtaining and responding to the result data of the image evaluation, constructing a screening model for the images in the album, that is, based on a deep learning network, by dividing the training set and the test set and optimizing the parameters, the images in the target image data set are quality evaluated and graded based on the corresponding weights, and the target screening image result data is obtained;

[0013] In response to the target filtering picture result data, the image data in the album is correspondingly filtered and visualized.

[0014] As a preferred solution of the present invention, the image screening influencing features include characteristic factors of a person's squinting and blinking state, a person's body shaking and trembling state, a person's face obstructing state, and a dark environment or exposure light state.

[0015] As a preferred solution of the present invention, wherein: based on a deep learning network, by dividing the training set and the test set and performing parameter optimization, the AIn-CNNs feature extraction model is specifically used to screen out the face and body area features of the character state and the ambient light state features of the character state and extract them, and then encode them in sequence to generate corresponding feature data of the face, body and ambient light state; and determine the target feature data based on the feature data.

[0016] As a preferred solution of the present invention, wherein: the face and body area images of the same person in the obtained target image data set are scaled to the same size;

[0017] Taking the center of the face and body area images as the origin, establish the image rectangular coordinate system;

[0018] Compare the feature data of each coordinate of the face and body area image with the target feature data, obtain the coordinates where the feature data differ, and determine them as the screened state, that is, the state of the person squinting and blinking, the state of the person shaking and trembling, the state of the person blocking the face, the state of the environment being dark or exposed to light;

[0019] By removing the corresponding coordinates in the face and body area images, the remaining image information of the face and body is filtered based on the target image data set and determined as result data of the image evaluation;

[0020] In response to the result data of the image evaluation, and based on the quality evaluation and grading, target screening picture result data is obtained, and the image data in the album is screened and visualized accordingly.

[0021] As a preferred solution of the present invention, wherein: based on a deep learning network, by dividing a training set and a test set and performing parameter optimization, it also includes generating abnormal samples of a target image data set in an image screening process and completing training;

[0022] The normal sample of the image screening and the sample to be screened of the target image are treated as a test sample pair for preprocessing operation;

[0023] The trained encoding module and decoding module are used to output the corresponding prediction results, and logical judgment is performed on the decoded results to perform real-time image screening and abnormal situation alarms.

[0024] As a preferred solution of the present invention, it also includes selecting people or scenery as the focus to increase the priority of being screened.

[0025] As a preferred solution of the present invention, wherein: based on the AI ​​big model, AI intelligent analysis is performed on the set of characteristic factors affecting image screening, and also includes optimizing the result data of image evaluation, specifically, the result data of image evaluation is processed in combination with fuzzy comprehensive evaluation to generate evaluated image optimization result data, that is, the image result data of AI intelligent analysis evaluation is evaluated using hesitant fuzzy sets to generate image screening optimization result data.

[0026] As a preferred solution of the present invention, the image evaluation result data is processed in combination with fuzzy comprehensive evaluation to generate evaluated image optimization result data, specifically as follows:

[0027] The hesitant fuzzy comprehensive evaluation determines the target image data set to be evaluated, and performs cluster analysis on the image screening influencing features and the corresponding image screening influencing feature factor set according to the image screening influencing features;

[0028] After determining the image screening influencing feature factor set and the image evaluation index, the corresponding evaluation level and judgment threshold are determined according to the image evaluation index;

[0029] According to different evaluation levels and judgment thresholds, hesitant fuzzy evaluation is performed in response to the result data of image evaluation to obtain the hesitant fuzzy evaluation matrix of the target image data set; specifically, let X = {x1, x2, …, x n} is the set of pictures in the target image dataset U = {y1,y2,…,y m} is a set of image evaluation indicators. According to the definitions of hesitant fuzzy sets and intuitionistic fuzzy sets, the comprehensive hesitant fuzzy matrix is ​​obtained through intelligent analysis of the AI ​​large model.

[0030] The comprehensive hesitant fuzzy evaluation matrix is ​​fuzzily synthesized with the weight vector, and the membership set is specifically calculated to obtain a generalized comparison table of the hesitant fuzzy set. After comparison with the characteristic function, the row sum, column sum, and row sum, column sum difference of the generalized comparison table are calculated to obtain the comprehensive evaluation result of the target image data set, and the image screening optimization result data is generated and output.

[0031] As a preferred solution of the present invention, wherein: the quality assessment and grading are performed, wherein the grading assessment calculation is shown as follows:

[0032] Among them, Q is the hierarchical evaluation value, Q1 is the evaluation value of the previous level, and E is the e-th level of the hierarchical evaluation. Evaluate the value in the neural network for the new layer.

[0033] In a second aspect, the present invention provides a system for an AI-based image evaluation and screening method, comprising:

[0034] An acquisition processing module is used to acquire image data in the album, perform feature extraction, cluster analysis and scene time recognition on the image data in sequence based on the classification features of the image data, and classify the image data into target image data of different groups according to the scene time classification features, and generate a corresponding target image data set;

[0035] An AI big model analysis module is used to call the target image data set through the AI ​​big model, classify, train, identify and associate the input target image data set, determine the image screening influencing features and the corresponding image screening influencing feature factor set in the image screening process, perform AI intelligent analysis on the image screening influencing feature factor set based on the AI ​​big model, and generate result data of image evaluation;

[0036] A grading evaluation module is used to obtain and respond to the result data of the image evaluation, and to construct a screening model for the images in the album, that is, based on a deep learning network, by dividing the training set and the test set and optimizing the parameters, and then performing quality evaluation and grading on the images in the target image data set based on the corresponding weights, and then obtaining the target screening image result data;

[0037] The output module is used to respond to the target screening picture result data and perform corresponding screening and visual presentation on the image data in the album.

[0038] Beneficial effects of the present invention:

[0039] The present invention applies AI to screen image quality, which can intelligently screen out more accurate, beautiful and visually appealing images, and combines hesitant fuzzy evaluation to improve the image grading evaluation method, further improving the hierarchical evaluation effect of image classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0041] Figure 1 It is a schematic diagram of the process of the present invention;

[0042] Figure 2 A flowchart of deep learning network processing in the method of the present invention;

[0043] Figure 3 Schematic diagram of the modular structure of the system of the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0045] Example 1

[0046] Reference Figure 1-Figure 2 , is an embodiment of the present invention, which provides an AI-based image evaluation and screening method, the method comprising:

[0047] First, the image data in the album is obtained, and based on the classification characteristics of the image data, the image data is subjected to feature extraction, cluster analysis and scene time recognition in sequence, and the image data is grouped according to the scene time classification characteristics into target image data of different groups, and a corresponding target image data set is generated.

[0048] Then, the target image data set is called by the AI ​​big model, and after data classification, training, recognition and association are performed on the input target image data set, the image screening influencing features and the corresponding image screening influencing feature factor set in the image screening process are determined, wherein the image screening influencing features include the characteristic factors of the person's squinting and blinking state, the person's body shaking state, the person's face blocking state, and the dark environment or the exposed light state;

[0049] Based on the deep learning network, by dividing the training set and the test set and optimizing the parameters, the AIn-CNNs feature extraction model is specifically used to screen the character state and environmental state in the target image data set to obtain the face and body area features of the character state and the ambient light state features and extract them, and then encode them in turn to generate the corresponding feature data of the face, body and ambient light state; the target feature data is determined according to the feature data.

[0050] Obtain the face and body area images of the same person in the target image data set and scale the images to the same size;

[0051] Taking the center of the face and body area images as the origin, establish the image rectangular coordinate system;

[0052] Compare the feature data of each coordinate of the face and body area image with the target feature data, obtain the coordinates where the feature data differ, and determine them as the screened state, that is, the state of the person squinting and blinking, the state of the person shaking and trembling, the state of the person blocking the face, the state of the environment being dark or exposed to light;

[0053] By removing the corresponding coordinates in the face and body area images, the remaining image information of the face and body is filtered based on the target image data set and determined as the result data of the image evaluation;

[0054] In response to the result data of image evaluation, and based on quality evaluation and grading, the target screening image result data is obtained, and the image data in the album is screened and visualized accordingly.

[0055] Based on the AI ​​big model, AI intelligent analysis is performed on the set of characteristic factors affecting image screening to generate result data for image evaluation.

[0056] At the same time, the result data of image evaluation is obtained and responded to, and a screening model for images in the album is constructed, that is, based on a deep learning network, after dividing the training set and the test set and optimizing the parameters, the images in the target image data set are evaluated and graded based on the corresponding weights, and the target screening image result data is obtained;

[0057] Finally, in response to the target filtering picture result data, the image data in the album is correspondingly filtered and visualized.

[0058] This embodiment further illustrates that based on the deep learning network, by dividing the training set and the test set and performing parameter optimization, it also includes generating abnormal samples of the target image data set in the image screening process and completing the training;

[0059] In this embodiment, preferably, the normal sample for image screening and the target image sample to be screened are treated as a test sample pair for preprocessing operation;

[0060] The trained encoding module and decoding module are used to output the corresponding prediction results, and logical judgment is performed on the decoded results to perform real-time image screening and abnormal situation alarms.

[0061] Preferably, this embodiment further includes selecting people or scenery as the focus to increase the priority of being screened.

[0062] Based on the above, this embodiment provides an AI-based image evaluation and screening method, which applies AI to perform image quality screening. After uploading a batch of images, such as photos taken after traveling, the images are first classified, and photos of the same scene are selected into the same group. At this time, the photos will be divided into several groups, and then the AI ​​large model is called to evaluate and grade the pictures in each group according to certain weights based on deep learning and neural networks. A certain person or scenery can be selected as the focus of attention. The evaluation and grading can be evaluated based on factors such as the squinting and blinking state of the person in the picture, the light state, the shaking state, etc., so as to achieve corresponding screening of the image data in the album.

[0063] Example 2

[0064] It should be emphasized in this embodiment that this embodiment performs optimization processing on the basis of embodiment 1, specifically, performs AI intelligent analysis on the image screening influencing feature factor set based on the AI ​​big model, and also includes optimizing the result data of image evaluation, specifically, processing the result data of image evaluation in combination with fuzzy comprehensive evaluation to generate the evaluated image optimization result data, that is, using hesitant fuzzy sets to evaluate the image result data of AI intelligent analysis evaluation, and generating image screening optimization result data. The image evaluation result data is processed in combination with fuzzy comprehensive evaluation to generate the evaluated image optimization result data, specifically as follows:

[0065] The hesitant fuzzy comprehensive evaluation determines the target image data set to be evaluated, and performs cluster analysis on the image screening influencing features and the corresponding image screening influencing feature factor set according to the image screening influencing features;

[0066] After determining the image screening influencing feature factor set and the image evaluation index, the corresponding evaluation level and judgment threshold are determined according to the image evaluation index;

[0067] According to different evaluation levels and judgment thresholds, hesitant fuzzy evaluation is performed in response to the result data of image evaluation to obtain the hesitant fuzzy evaluation matrix of the target image data set; specifically, let X = {x1, x2, …, x n} is the set of pictures in the target image dataset U = {y1,y2,…,y m} is a set of image evaluation indicators. According to the definitions of hesitant fuzzy sets and intuitionistic fuzzy sets, the comprehensive hesitant fuzzy matrix is ​​obtained through intelligent analysis of the AI ​​large model.

[0068] The comprehensive hesitant fuzzy evaluation matrix and the weight vector are fuzzy synthesized, and the membership set is specifically calculated to obtain the generalized comparison table of the hesitant fuzzy set. After comparison with the generalized comparison table using the characteristic function, the row sum, column sum, and row sum and column sum difference are calculated to obtain the comprehensive evaluation result of the target image data set, and the image screening optimization result data is generated and output;

[0069] The characteristic function is defined as follows:

[0070] Among them, v and s are the row value and column value in the generalized comparison table respectively;

[0071] The calculation of the image screening optimization result data is shown in the following formula:

[0072]

[0073] Among them, S i (X1), S i(X2) represents the membership score and non-membership score of the i-th image in the target image dataset respectively; is the comprehensive evaluation score of the target image dataset. Sorting is performed to obtain the filtered images in the target image data set to form image optimization result data.

[0074] In addition, this embodiment performs quality assessment and grading, wherein the grading assessment calculation is shown in the following formula:

[0075] Among them, Q is the hierarchical evaluation value, Q1 is the evaluation value of the previous level, and E is the e-th level of the hierarchical evaluation. Evaluate the value in the neural network for the new layer.

[0076] Based on the above, it can be seen that this embodiment, on the basis of implementation 1, adds a hesitant evaluation of the quality evaluation process of the images in the hierarchical process in order to prevent the uncertainty and poor evaluation effect in the hierarchical quality evaluation process of the images, thereby effectively achieving a more prepared judgment and evaluation after image grading, which is conducive to accurately screening out poor images, and can intelligently screen out more accurate, beautiful and visually appealing images. The image grading evaluation method is improved by combining the hesitant fuzzy evaluation, and the hierarchical judgment effect of image classification is further improved.

[0077] Example 3

[0078] Reference Figure 3 This embodiment provides a system for an AI-based image evaluation and screening method, including:

[0079] An acquisition processing module is used to acquire image data in the album, perform feature extraction, cluster analysis and scene time recognition on the image data in sequence based on the classification features of the image data, and classify the image data into target image data of different groups according to the scene time classification features, and generate a corresponding target image data set;

[0080] The AI ​​big model analysis module is used to call the target image data set through the AI ​​big model, classify, train, identify and associate the input target image data set, determine the image screening influencing features and the corresponding image screening influencing feature factor set in the image screening process, perform AI intelligent analysis on the image screening influencing feature factor set based on the AI ​​big model, and generate result data of image evaluation;

[0081] The grading evaluation module is used to obtain and respond to the result data of image evaluation, and to build a screening model for images in the album. That is, based on the deep learning network, after dividing the training set and the test set and optimizing the parameters, the images in the target image data set are evaluated and graded based on the corresponding weights, and the target screening image result data is obtained;

[0082] The output module is used to respond to the target filtering image result data and perform corresponding filtering and visual presentation on the image data in the album.

[0083] In summary, the present invention applies AI to screen image quality, which can intelligently screen out more accurate, beautiful and visually appealing images, and combines hesitant fuzzy evaluation to improve the image grading evaluation method, further improving the hierarchical evaluation effect of image classification.

[0084] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.

[0085] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0086] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0087] Any process or method description in the flow chart or otherwise described herein can be understood to represent a module, fragment or portion of a code including one or more executable instructions for implementing the steps of a specific logical function or process. And the scope of the preferred embodiment of the present application includes other implementations, in which the functions may not be performed in the order shown or discussed, including in a substantially simultaneous manner or in a reverse order according to the functions involved.

[0088] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, which can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with these instruction execution systems, apparatuses or devices.

[0089] It should be understood that the various parts of the present application can be implemented with hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented with software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the above embodiment method can be completed by instructing the relevant hardware through a program, which can be stored in a computer-readable storage medium, and when the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0090] In addition, each functional unit in each embodiment of the present application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into one module. The above-mentioned integrated module can be implemented in the form of hardware or in the form of a software functional module. If the above-mentioned integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. The storage medium can be a read-only memory, a disk or an optical disk, etc.

[0091] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An AI-based image evaluation and screening method, characterized in that: include: Acquire image data in the photo album, perform feature extraction, cluster analysis and scene time recognition on the image data in sequence based on classification features of the image data, group the image data into target image data of different groups according to the scene time classification features, and generate corresponding target image data sets; The target image data set is called by the AI ​​big model, and after data classification, training, recognition and association are performed on the input target image data set, the image screening influencing features and the corresponding image screening influencing feature factor set in the image screening process are determined, and the image screening influencing feature factor set is subjected to AI intelligent analysis based on the AI ​​big model to generate result data of image evaluation; Obtaining and responding to the result data of the image evaluation, constructing a screening model for the images in the album, that is, based on a deep learning network, by dividing the training set and the test set and optimizing the parameters, the images in the target image data set are quality evaluated and graded based on the corresponding weights, and the target screening image result data is obtained; In response to the target screening picture result data, the image data in the album is screened and visualized accordingly; The image screening influencing features include the characteristic factors of the person's squinting and blinking state, the person's body shaking and trembling state, the person's face blocking state, and the dark environment or exposure light state; Based on the deep learning network, by dividing the training set and the test set and optimizing the parameters, the AIn-CNNs feature extraction model is specifically used to screen the human state and environmental state in the target image data set to obtain the facial and body area features of the human state and the environmental light state features, and then encode and generate the corresponding facial, body and environmental light state feature data in sequence; and determine the target feature data according to the feature data; Obtain the face and body area images of the same person in the target image data set and scale the images to the same size; Taking the center of the face and body area images as the origin, establish the image rectangular coordinate system; Compare the feature data of each coordinate of the face and body area image with the target feature data, obtain the coordinates where the feature data differ, and determine them as the screened state, that is, the state of the person squinting and blinking, the state of the person shaking and trembling, the state of the person blocking the face, the state of the environment being dark or exposed to light; By removing the corresponding coordinates in the face and body area images, the remaining image information of the face and body is filtered based on the target image data set and determined as result data of the image evaluation; In response to the result data of the image evaluation, and based on the quality evaluation and grading, target screening picture result data is obtained, and the image data in the album is screened and visualized accordingly.

2. The AI-based image evaluation and screening method according to claim 1, characterized in that: Based on the deep learning network, by dividing the training set and the test set and optimizing the parameters, it also includes generating abnormal samples of the target image data set in the image screening process and completing the training; The normal sample of the image screening and the sample to be screened of the target image are treated as a test sample pair for preprocessing operation; The trained encoding module and decoding module are used to output the corresponding prediction results, and logical judgment is performed on the decoded results to perform real-time image screening and abnormal situation alarms.

3. The AI-based image evaluation and screening method according to claim 1, characterized in that: It also includes selecting people or scenery as the focus to increase the priority of being screened.

4. The AI-based image evaluation and screening method according to claim 1, characterized in that: Based on the AI ​​big model, AI intelligent analysis is performed on the set of characteristic factors affecting image screening, and the result data of image evaluation is also optimized. Specifically, the result data of image evaluation is combined with fuzzy comprehensive evaluation for processing to generate evaluated image optimization result data, that is, the image result data of AI intelligent analysis evaluation is evaluated using hesitant fuzzy sets to generate image screening optimization result data.

5. The AI-based image evaluation and screening method according to claim 4, characterized in that: The image evaluation result data is processed in combination with the fuzzy comprehensive evaluation to generate the evaluated image optimization result data, specifically as follows: The hesitant fuzzy comprehensive evaluation determines the target image data set to be evaluated, and performs cluster analysis on the image screening influencing features and the corresponding image screening influencing feature factor set according to the image screening influencing features; After determining the image screening influencing feature factor set and the image evaluation index, the corresponding evaluation level and judgment threshold are determined according to the image evaluation index; According to different evaluation levels and judgment thresholds, and in response to the result data of image evaluation, a hesitant fuzzy evaluation is performed to obtain the hesitant fuzzy evaluation matrix of the target image data set; specifically, a set of pictures in the target image data set and a set of image evaluation indicators are set. According to the definitions of hesitant fuzzy sets and intuitionistic fuzzy sets, a comprehensive hesitant fuzzy matrix is ​​obtained through intelligent analysis of the AI ​​large model. ; The comprehensive hesitant fuzzy matrix and the weight vector are fuzzy synthesized, and the membership set is specifically calculated to obtain a generalized comparison table of the hesitant fuzzy set. After comparison with the characteristic function, the row sum, column sum and row sum, column sum difference of the generalized comparison table are calculated to obtain the comprehensive evaluation result of the target image data set, and the image screening optimization result data is generated and output.

6. The AI-based image evaluation and screening method according to claim 5, characterized in that: The quality assessment and grading are performed, wherein the grading assessment calculation is shown as follows: ； in, is the graded evaluation value, is the evaluation value of the previous level, E For the graded assessment e Number of layers, Evaluate the value in the neural network for the new layer.

7. A system applied to the AI-based image evaluation and screening method as claimed in claim 1, characterized in that: include: An acquisition processing module is used to acquire image data in the album, perform feature extraction, cluster analysis and scene time recognition on the image data in sequence based on the classification features of the image data, and classify the image data into target image data of different groups according to the scene time classification features, and generate a corresponding target image data set; An AI big model analysis module is used to call the target image data set through the AI ​​big model, classify, train, identify and associate the input target image data set, determine the image screening influencing features and the corresponding image screening influencing feature factor set in the image screening process, perform AI intelligent analysis on the image screening influencing feature factor set based on the AI ​​big model, and generate result data of image evaluation; A grading evaluation module is used to obtain and respond to the result data of the image evaluation, and to construct a screening model for the images in the album, that is, based on a deep learning network, by dividing the training set and the test set and optimizing the parameters, and then performing quality evaluation and grading on the images in the target image data set based on the corresponding weights, and then obtaining the target screening image result data; The output module is used to respond to the target screening picture result data and perform corresponding screening and visual presentation on the image data in the album.

Images