Image intelligent scoring and ranking method, system and storage medium based on element recognition

Abstract

The application discloses an image intelligent scoring and sorting method and system based on element identification and a storage medium, relates to image recognition technology, and comprises the following steps: acquiring a plurality of images to be scored and a user feature vector; identifying constituent elements from the images, generating a first vector according to the constituent elements identified from the images; analyzing color composition from the images, generating a second vector according to the color composition; splicing the first vector and the second vector to obtain a third vector; scoring the images according to the similarity between the third vector and the user feature vector; and sorting the plurality of scored images. Through the scheme, the images can be intelligently scored, the images of interest of customers can be evaluated, and intelligent sorting can be realized to meet application needs.

Description

Technical Field

[0001] The present application relates to image recognition technology, and in particular to an image intelligent scoring and sorting method, system and storage medium based on element recognition. Background Art

[0002] When pushing ads to smart devices, ads are generally pushed based on user characteristics. Current technologies primarily focus on analyzing which types of ads (and therefore, which products) users are interested in, without distinguishing between the types of products users are interested in. For example, the same product may feature multiple advertising posters, and users may have different preferences for different posters. Posters that appeal to users more can attract users to stay and increase the likelihood of conversions.

[0003] Therefore, when the same product has multiple product posters, how to push posters to users becomes an issue that needs to be studied. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes an image intelligent scoring and ranking method, system and storage medium based on element recognition to intelligently score posters.

[0005] In one aspect, an embodiment of the present application provides an image intelligent scoring and ranking method based on element recognition, comprising the following steps:

[0006] Obtain multiple images to be rated and user feature vectors;

[0007] Identifying constituent elements from the image, and generating a first vector based on the constituent elements identified from the image to represent element features of the image;

[0008] Analyzing color composition from the image, and generating a second vector based on the color composition to represent color features of the image;

[0009] concatenating the first vector and the second vector to obtain a third vector for representing features of the image;

[0010] Scoring the image according to the similarity between the third vector and the user feature vector;

[0011] Sort the scored images.

[0012] In some embodiments, the scoring of the image according to the similarity between the third vector and the user feature vector is specifically:

[0013] Calculating the Euclidean distance between the third vector and the user feature vector;

[0014] Get the correspondence table between Euclidean distance and score;

[0015] The score is determined based on the calculated Euclidean distance and the corresponding relationship table.

[0016] In some embodiments, identifying constituent elements from the image and generating a first vector based on the constituent elements identified from the image to represent element features of the image specifically includes:

[0017] Identify multiple constituent elements from the image, remove product objects from the constituent elements, and obtain remaining constituent elements;

[0018] The first vector is constructed according to the characteristic vectors of the constituent elements.

[0019] In some embodiments, the first vector is constructed based on the characteristic vectors of the constituent elements, specifically:

[0020] The eigenvectors are averaged to obtain a first vector.

[0021] In some embodiments, analyzing color composition from the image and generating a second vector based on the color composition specifically includes:

[0022] Determine the top three color vectors in the image;

[0023] The top three color vectors are weighted to obtain the second vector.

[0024] In some embodiments, the plurality of images belong to different posters containing an image of the same product.

[0025] In some embodiments, the user feature vector is determined based on user historical behavior.

[0026] On the other hand, the embodiment of the present application discloses an image intelligent scoring and ranking system based on element recognition, comprising:

[0027] An acquisition unit, configured to acquire a plurality of images to be rated and user feature vectors;

[0028] a constituent element extraction unit, configured to identify constituent elements from the image, and generate a first vector based on the constituent elements identified from the image, so as to represent element features of the image;

[0029] a color feature extraction unit, configured to analyze color composition from the image and generate a second vector according to the color composition to represent color features of the image;

[0030] a concatenation unit, configured to concatenate the first vector and the second vector to obtain a third vector for representing a feature of the image;

[0031] a scoring unit, configured to score the image based on the similarity between the third vector and the user feature vector;

[0032] The sorting unit is used to sort the multiple scored images.

[0033] On the other hand, the embodiment of the present application discloses an image intelligent scoring and ranking system based on element recognition, comprising:

[0034] Memory, used to store programs;

[0035] A processor is used to load the program to execute the image intelligent scoring and sorting method based on element recognition.

[0036] On the other hand, an embodiment of the present application discloses a computer-readable storage medium storing a program, which, when executed by a processor, implements the image intelligent scoring and sorting method based on element recognition.

[0037] This embodiment of the application obtains multiple images to be scored and user feature vectors, analyzes the image's color and composition to generate a first vector and a second vector, then concatenates the first and second vectors to generate a third vector to characterize the image's features. The images are then scored based on the similarity between the third vector and the user feature vector, and the scored images are sorted. This approach allows the user's preferences, such as the basic element composition and color composition of the image, to be compared to identify images that the user may be more interested in, thereby optimizing advertising delivery. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0039] Figure 1 This is a flowchart of an image intelligent scoring and ranking method based on element recognition provided by an embodiment of the present application;

[0040] Figure 2 This is a block diagram of an image intelligent scoring and ranking system based on element recognition provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] To make the objectives, technical solutions, and advantages of this application more clear, the following will refer to the drawings in the embodiments of this application to clearly and completely describe the technical solutions of this application through implementation methods. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0042] In the description of the present invention, "several" means more than one, "plurality" means more than two, "greater than," "less than," and "exceed" are understood to exclude the number itself, while "above," "below," and "within" are understood to include the number itself. The use of "first" and "second" in the description is solely for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance, implicitly specifying the number of the indicated technical features, or implicitly specifying the order of the indicated technical features.

[0043] In the description of the present invention, unless otherwise clearly defined, words such as “setting” should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meaning of the above words in the present invention based on the specific content of the technical solution.

[0044] In the description of the present invention, reference to terms such as "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that a specific feature or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0045] Reference Figure 1 , an image intelligent scoring and sorting method based on element recognition, which can be applied to advertising push, wallpaper recommendation, etc. of smart devices. The embodiment of this application can be applied to an all-in-one educational machine to realize intelligent wallpaper (advertisement) push.

[0046] The method comprises the following steps:

[0047] S1. Obtain multiple images to be rated and user feature vectors. The user feature vectors are determined based on the user's historical behavior.

[0048] In this embodiment, images to be rated generally refer to wallpapers / advertising posters, etc. that need to be rated and sorted. A user feature vector refers to a vector used to describe a user's preferences for image colors and items. The user feature vector can be learned from the user's historical records. For example, based on the user's settings of themes, wallpaper colors, and other elements. Alternatively, the user's preferred elements can be determined through the user's browsing behavior. For example: cats, dogs, cars, airplanes, etc. These features can be learned from the user's browsing history across various channels. In this embodiment, the user feature vector consists of a preferred element part and a preferred color part. The user's preferred elements and preferred colors can be learned separately. Then, a component element vector and a color vector representing the user's preferences are generated separately, and the two are spliced ​​together to form the user feature vector.

[0049] S2. Identify constituent elements from the image, and generate a first vector based on the constituent elements identified from the image to represent element features of the image.

[0050] In this embodiment, a general image recognition model can be used to first identify the elements in the model, such as cars, airplanes, cats, dogs, and people. Representations of these objects in the model are then searched. The following processing methods can then be used to obtain the first vector.

[0051] Method 1: The first vector can be represented by averaging the vectors of the constituent elements.

[0052] The second method is to weight the vector according to the number of times its constituent elements appear.

[0053] Method three considers the area of ​​the elements and constructs the first N elements by area. This can be done by averaging, weighting by area, or concatenating, where N is a positive integer. If concatenating is used, the compatibility of the vector length with the user's feature vector needs to be considered.

[0054] S3. Analyze color composition from the image, and generate a second vector based on the color composition to represent color features of the image.

[0055] In this embodiment, processing can be performed based on the color ratio. To reduce the number of element statistics, the color statistics can be divided according to a certain color width. For example, if there are a total of consecutive numbers 0 to 65536, 64 can be used as a unit, and all colors within a color range with a width of 64 can be assigned to a certain color. Reducing the number of colors can reduce model complexity and prediction difficulty.

[0056] In some embodiments, analyzing color composition from the image and generating a second vector based on the color composition specifically includes:

[0057] Determine the top three color vectors in the image;

[0058] The top three color vectors are weighted to obtain the second vector.

[0059] In this way, the main colors can be separated, and then the three color vectors are weighted to obtain the second vector, and the weighted weight is determined according to the pixels occupied by the color.

[0060] S4. Concatenate the first vector and the second vector to obtain a third vector for representing features of the image.

[0061] In this step, the first and second vectors are directly concatenated. This concatenation creates an X+Y-dimensional third vector, representing the image's features. Essentially, this third vector describes the image's components and color characteristics in the semantic space of the vectors.

[0062] S5. Score the image based on the similarity between the third vector and the user feature vector. In this embodiment, the closer the two vectors are in the semantic space, the higher the similarity between the two vectors. By comparing the similarity between the two vectors, the degree of overlap between the user's preferences and the image can be determined. Therefore, the image can be scored based on the similarity. Generally, the similarity between two vectors can be determined using methods such as Euclidean distance. The smaller the distance, the higher the score. Here, the distance value range can be divided and corresponding to a certain score.

[0063] Specifically, the scoring of the image according to the similarity between the third vector and the user feature vector is specifically: calculating the Euclidean distance between the third vector and the user feature vector; obtaining a correspondence table between the Euclidean distance and the score; and determining the score based on the calculated Euclidean distance and the correspondence table.

[0064] S6. Sort the scored images.

[0065] This step sorts the scored images according to the scores. Applications after sorting include, but are not limited to, recommending a specific number of top-ranked images to users, displaying them according to the sorting results, setting the exposure time of images according to the sorting, etc.

[0066] This embodiment of the application obtains multiple images to be scored and user feature vectors, analyzes the image's color and composition to generate a first vector and a second vector, then concatenates the first and second vectors to generate a third vector to characterize the image's features. The images are then scored based on the similarity between the third vector and the user feature vector, and the scored images are sorted. This approach allows the user's preferences, such as the basic element composition and color composition of the image, to be compared to identify images that the user may be more interested in, thereby optimizing advertising delivery.

[0067] In some embodiments, since ad type matching is performed by another model, identifying the products that users may be interested in is not the task of this solution. This model is mainly used to sort different posters of the same product, so when processing, the corresponding products in the component elements can be eliminated.

[0068] Specifically, identifying constituent elements from the image and generating a first vector based on the constituent elements identified from the image to represent element features of the image specifically includes:

[0069] Identify multiple constituent elements from the image, remove product objects from the constituent elements, and obtain remaining constituent elements;

[0070] The first vector is constructed according to the characteristic vectors of the constituent elements.

[0071] In this way, the influence of product objects on poster features can be reduced, allowing the model to focus more on other elements of the poster, thereby increasing the accuracy of model predictions.

[0072] Reference Figure 2 The present application discloses an intelligent image scoring and ranking system based on element recognition, including:

[0073] An acquisition unit, configured to acquire a plurality of images to be rated and user feature vectors;

[0074] a constituent element extraction unit, configured to identify constituent elements from the image, and generate a first vector based on the constituent elements identified from the image, so as to represent element features of the image;

[0075] a color feature extraction unit, configured to analyze color composition from the image and generate a second vector according to the color composition to represent color features of the image;

[0076] a concatenation unit, configured to concatenate the first vector and the second vector to obtain a third vector for representing a feature of the image;

[0077] a scoring unit, configured to score the image based on the similarity between the third vector and the user feature vector;

[0078] The sorting unit is used to sort the multiple scored images.

[0079] On the other hand, the embodiment of the present application discloses an image intelligent scoring and ranking system based on element recognition, comprising:

[0080] Memory, used to store programs;

[0081] A processor is used to load the program to execute the image intelligent scoring and sorting method based on element recognition.

[0082] On the other hand, an embodiment of the present application discloses a computer-readable storage medium storing a program, which, when executed by a processor, implements the image intelligent scoring and sorting method based on element recognition.

[0083] If the integrated unit described in this application is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of this application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0084] Note that the above are only preferred embodiments of the present application and the technical principles employed. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments and may include many other equivalent embodiments without departing from the scope of the present application. The scope of the present application is determined by the scope of the appended claims.

Claims

1. An image intelligent scoring and ranking method based on element recognition, characterized in that: The following steps are involved: Obtain multiple images to be rated and user feature vectors; Identifying constituent elements from the image, and generating a first vector based on the constituent elements identified from the image to represent element features of the image; Analyzing color composition from the image, and generating a second vector based on the color composition to represent color features of the image; concatenating the first vector and the second vector to obtain a third vector for representing features of the image; Scoring the image according to the similarity between the third vector and the user feature vector; Sort the scored images; The identifying constituent elements from the image and generating a first vector according to the constituent elements identified from the image to represent element features of the image specifically includes: Identify multiple constituent elements from the image, remove product objects from the constituent elements, and obtain remaining constituent elements; constructing the first vector according to the characteristic vectors of the constituent elements; Analyzing color composition from the image and generating a second vector according to the color composition specifically includes: Determine the top three color vectors in the image; The top three color vectors are weighted to obtain the second vector.

2. The method for intelligent image scoring and ranking based on element recognition according to claim 1, characterized in that: Scoring the image according to the similarity between the third vector and the user feature vector is specifically: Calculating the Euclidean distance between the third vector and the user feature vector; Get the correspondence table between Euclidean distance and score; The score is determined based on the calculated Euclidean distance and the corresponding relationship table.

3. The image intelligent scoring and ranking method based on element recognition according to claim 1 is characterized in that: The first vector is constructed according to the characteristic vectors of the constituent elements, specifically: The eigenvectors are averaged to obtain a first vector.

4. The method for intelligent image scoring and ranking based on element recognition according to claim 1, characterized in that: The plurality of images belong to different posters containing an image of the same product.

5. The method for intelligent image scoring and ranking based on element recognition according to claim 1, characterized in that: The user feature vector is determined based on the user's historical behavior.

6. An intelligent image scoring and ranking system based on element recognition, characterized in that: include: An acquisition unit, configured to acquire a plurality of images to be rated and user feature vectors; a constituent element extraction unit, configured to identify constituent elements from the image, and generate a first vector based on the constituent elements identified from the image, so as to represent element features of the image; a color feature extraction unit, configured to analyze color composition from the image and generate a second vector according to the color composition to represent color features of the image; a concatenation unit, configured to concatenate the first vector and the second vector to obtain a third vector for representing a feature of the image; a scoring unit, configured to score the image based on the similarity between the third vector and the user feature vector; The sorting unit is used to sort the multiple scored images.

7. An image intelligent scoring and ranking system based on element recognition, characterized in that: include: Memory, used to store programs; A processor, configured to load the program to execute the image intelligent scoring and ranking method based on element recognition as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that It stores a program, and when the program is executed by a processor, it implements the image intelligent scoring and sorting method based on element recognition as described in any one of claims 1 to 5.

Images