Image real-name conversion method, device and electronic device based on artificial intelligence

Through the image real-name method based on artificial intelligence, the average feature and target feature of the image are used to select candidate reference images, which solves the problem of deviation and low efficiency of the image real-name result, and achieves higher accuracy and efficiency.

CN114581989BActive Publication Date: 2024-12-03TANGSHAN XINDU TECHNOLOGY DEVELOPMENT CO LTD
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
CN202210217981.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-07
Publication Date
2024-12-03
Estimated Expiration
2042-03-07

AI Technical Summary

Technical Problem

There are deviations in the results of image realization and low efficiency, especially in the fields of computer vision and public safety, which are difficult to effectively solve in the existing technology.

Method used

Using an image real-name method based on artificial intelligence, by determining the average feature of the image and the features of the target image, selecting candidate reference images, and determining real-name information based on these features, improving the accuracy and efficiency of real-nameization.

Benefits of technology

It effectively improves the accuracy and efficiency of image realization, reduces the impact of average feature deviation on realization results, and enhances the ability to support image realization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114581989B_ABST
    Figure CN114581989B_ABST
Patent Text Reader

Abstract

The present disclosure provides an artificial intelligence-based image real-name method, device and electronic device, which relates to the field of computer vision and image processing technology, and can be applied to scenarios such as smart cloud, cloud computing, smart city, security and urban governance. The specific implementation scheme includes: for M images to be real-named, determining the first average image feature associated with the M images, M is an integer greater than 1; for N target images in the M images, determining the first image feature associated with each of the N target images, the quality assessment value of the target image is greater than a preset threshold, N is an integer greater than 1 and less than M; determining at least one candidate reference image based on the first average image feature and the first image feature associated with each target image; and determining the real-name information for the M images based on the object identity feature indicated by at least one candidate reference image.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of computer vision and image processing technology, and can be applied to scenarios such as smart cloud, cloud computing, smart city, security and urban governance. Background Art

[0002] Image real-name registration is of great significance to computer vision, public security and other fields, and can provide reliable data support for determining the name, category and other attribute information of unknown data. However, in some scenarios, the results of image real-name registration are biased and the efficiency of real-name registration is low. Summary of the invention

[0003] The present invention provides an artificial intelligence-based image real-name conversion method, device, and electronic device.

[0004] According to one aspect of the present disclosure, a method for image real-name authentication is provided, comprising: for M images to be authenticated, determining a first average image feature associated with the M images, where M is an integer greater than 1; for N target images among the M images, determining a first image feature associated with each of the N target images, wherein a quality assessment value of the target image is greater than a preset threshold, where N is an integer greater than 1 and less than M; determining at least one candidate reference image based on the first average image feature and the first image feature associated with each of the target images; and determining real-name information for the M images based on an object identity feature indicated by the at least one candidate reference image.

[0005] According to one aspect of the present disclosure, there is provided an image real-name conversion device, comprising: a first determination module, configured to determine, for M images to be real-named, a first average image feature associated with the M images, where M is an integer greater than 1; a second determination module, configured to determine, for N target images among the M images, a first image feature associated with each of the N target images, wherein a quality assessment value of the target image is greater than a preset threshold, where N is an integer greater than 1 and less than M; a third determination module, configured to determine at least one candidate reference image based on the first average image feature and the first image feature associated with each of the target images; and a fourth determination module, configured to determine the real-name information for the M images based on the object identity feature indicated by the at least one candidate reference image.

[0006] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor and a memory in communication with the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the above-mentioned image real-name conversion method.

[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the above-mentioned image real-name conversion method.

[0008] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, wherein the computer program implements the above-mentioned image real-name conversion method when executed by a processor.

[0009] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0011] Figure 1 The system architecture of the image real-name conversion method and device according to an embodiment of the present disclosure is schematically shown;

[0012] Figure 2 A flowchart of an image real-name conversion method according to an embodiment of the present disclosure is schematically shown;

[0013] Figure 3 A flowchart of an image real-name conversion method according to another embodiment of the present disclosure is schematically shown;

[0014] Figure 4 A schematic diagram schematically shows a process of image real-name conversion according to an embodiment of the present disclosure;

[0015] Figure 5 A block diagram of an image real-name conversion device according to an embodiment of the present disclosure is schematically shown;

[0016] Figure 6 A block diagram of an electronic device for executing an image real-name conversion method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0017] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0018] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0019] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0020] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0021] The embodiment of the present disclosure provides an artificial intelligence-based image real-name method. The method includes: for M images to be real-named, determining the first average image feature associated with the M images, where M is an integer greater than 1; for N target images among the M images, determining the first image feature associated with each of the N target images, where the quality assessment value of the target image is greater than a preset threshold, where N is an integer greater than 1 and less than M; determining at least one candidate reference image based on the first average image feature and the first image feature associated with each target image; and determining the real-name information for the M images based on the object identity feature indicated by the at least one candidate reference image.

[0022] Figure 1 The system architecture of the image real-name method and device according to an embodiment of the present disclosure is schematically shown. It should be noted that: Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0023] The system architecture 100 according to this embodiment may include data acquisition terminals 101, 102, a network 103 and a server 104. The network 103 is used to provide a medium for a communication link between the data acquisition terminals 101, 102 and the server 104. The network 103 may include various connection types, such as wired or wireless communication links or optical fiber cables, etc. The server 104 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud computing, network services, and middleware services.

[0024] The data acquisition terminals 101 and 102 interact with the server 104 through the network 103 to receive or send data, etc. The data acquisition terminals 101 and 102 can be used to collect image information of the environment, such as face information of a specific area, etc. The data acquisition terminals 101 and 102 can also be used to collect other picture information, such as picture information on documents, such as portrait information contained in identity cards, driver's licenses, etc. The data acquisition terminals 101 and 102 can be, for example, cameras, video cameras, image scanning devices, etc.

[0025] The server 104 may be a server that provides various services, for example, it may be a background processing server that performs image real-name registration based on the image information or face information provided by the data acquisition terminals 101 and 102 (only as an example).

[0026] For example, the server 104 responds to image information obtained from the data acquisition terminals 101 and 102, where the image information includes M images, where the M images may be, for example, a set of images belonging to the same object after clustering. For the M images, a first average image feature associated with the M images is determined, N target images are obtained from the M images, a first image feature associated with each of the N target images is determined, at least one candidate reference image is determined based on the first average image feature and the first image feature associated with each target image, and real-name information for the M images is determined based on the object identity feature indicated by the at least one candidate reference image.

[0027] It should be noted that the image real-name method provided in the embodiment of the present disclosure can be executed by the server 104. Accordingly, the image real-name device provided in the embodiment of the present disclosure can be set in the server 104. The image real-name method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 104 and can communicate with the data acquisition terminals 101, 102 and / or the server 104. Accordingly, the image real-name device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 104 and can communicate with the data acquisition terminals 101, 102 and / or the server 104.

[0028] It should be understood that Figure 1 The numbers of data collection terminals, networks and servers in the embodiment are only for illustration purposes. Any number of data collection terminals, networks and servers may be provided according to implementation requirements.

[0029] The present disclosure provides an artificial intelligence-based image real-name method. Figure 1 The system architecture of Figures 2 to 4 The image real-name method according to the exemplary embodiment of the present disclosure is described as follows. The image real-name method according to the exemplary embodiment of the present disclosure can be, for example, Figure 1 The server 104 is shown to execute.

[0030] Figure 2 The flowchart of the method for image real-name conversion according to an embodiment of the present disclosure is schematically shown.

[0031] like Figure 2 As shown, the image real-name method 200 of the embodiment of the present disclosure may include, for example, operations S210 to S240.

[0032] In operation S210 , for M images to be authenticated, a first average image feature associated with the M images is determined, where M is an integer greater than 1.

[0033] In operation S220, for N target images among the M images, a first image feature associated with each of the N target images is determined, and a quality assessment value of the target image is greater than a preset threshold, where N is an integer greater than 1 and less than M.

[0034] In operation S230, at least one candidate reference image is determined based on the first average image feature and the first image feature associated with each target image.

[0035] In operation S240, real name information for the M images is determined based on the object identity feature indicated by the at least one candidate reference image.

[0036] The following is an example flow of each operation of the image real-name conversion method of this embodiment.

[0037] Exemplarily, the image data involved in this embodiment can be obtained through various public, legal and compliant methods, such as from a public data set, or obtained by a data acquisition terminal after obtaining authorization from a user associated with the image data. Image data is not scene data for a specific user and cannot reflect the personal information of a specific user. The application scope of image data is limited to the scope that the user has the right to know and has authorized to use.

[0038] For the M images to be authenticated, a first average image feature associated with the M images is determined, where M is an integer greater than 1. The M images to be authenticated may include, for example, multiple images of the same object. The M images may be acquired, for example, by an imaging element of a data acquisition terminal, or may be acquired from other databases by the data acquisition terminal.

[0039] The M images may include, for example, a face image taken by an image acquisition device, or an image containing a face acquired based on an ID card, or other data containing features of a face image.

[0040] The first average image feature associated with the M images can be obtained based on the first image feature of each of the M images. For example, the first image feature of each of the M images is determined, and the image feature mean of the first image features of the M images is used as the first average image feature associated with the M images.

[0041] The first image feature may be, for example, an image feature vector obtained by extracting image features. The first average image feature may be, for example, an average value calculated from the first image feature vectors of M images, and the first average image feature indicates the centroid of the image set or category cluster composed of the M images.

[0042] For N target images among the M images, a first image feature associated with each of the N target images is determined, and a quality assessment value of the target image is greater than a preset threshold, where N is an integer greater than 1 and less than M.

[0043] Exemplarily, selecting N target images from M images may include, for example, selecting, based on the quality evaluation value of each image in the M images, an image corresponding to a quality evaluation value greater than a preset threshold as a target image, so as to obtain N target images from the M images.

[0044] After the acquired image is recognized and analyzed, the image quality can be evaluated according to the image characteristics such as the shooting angle, clarity, sensitivity, light, blur, image size, grayscale, etc., to obtain a quality evaluation value for the image. If the quality evaluation value is higher than the preset threshold, it indicates that the image quality is good, and if the quality evaluation value is lower than the preset threshold, it indicates that the image quality is poor.

[0045] A quality evaluation value ranking of the M images may be generated according to the quality evaluation value of each image in the M images, and the first N images may be determined as target images according to the quality evaluation value ranking. An image having a quality evaluation value greater than a preset threshold may also be selected from the M images as the target image. Since each of the M images has a first image feature, after determining N target images from the M images, the first image feature associated with each of the N target images is obtained.

[0046] At least one candidate reference image is determined based on the first average image feature and the first image feature associated with each target image. The candidate reference image may be determined, for example, from the candidate images. The candidate images may include, for example, images in a database with object identity information, for example, the database with object identity information may include a database of facial image features exceeding a preset threshold.

[0047] At least one candidate reference image is determined based on the first average image feature and the first image feature of each target image. Exemplarily, at least one candidate reference image can be determined from the candidate images based on the first average image feature, or at least one candidate reference image can be determined from the candidate images based on the first image feature of each target image.

[0048] Exemplarily, a second image feature for each image in the candidate images is obtained. At least one first candidate reference image is determined from the candidate images based on the first average image feature and the second image feature. At least one second candidate reference image is determined from the candidate images based on the second image feature and the first image feature associated with each target image. At least one candidate reference image is determined based on at least one first candidate reference image and at least one second candidate reference image.

[0049] Through the embodiments of the present disclosure, a candidate reference image in the candidate images is determined based on the first average image feature and the first image feature associated with each target image. When there is a deviation between the first average image feature and the object identity feature of the M images, the candidate reference image in the candidate images is determined by the first image feature associated with each target image, which can effectively eliminate the influence of the deviation of the first average image feature. By increasing the number of candidate reference images obtained from the candidate images, the accuracy of real-name verification can be improved.

[0050] The candidate reference image may be obtained from the candidate image, for example. Each of the candidate images has a second image feature, and each second image feature may reflect each image feature in the candidate image, for example. Each of the candidate images may also include an object identity feature associated with the image, for example. The object identity feature may include information such as the name and attributes of the object indicated by each candidate image.

[0051] The first candidate reference image and the second candidate reference image can be selected and determined from the candidate images, for example. At least one first candidate reference image is determined from the candidate images according to the first average image feature and the second image feature. For example, according to the similarity between the first average image feature and the second image feature of each candidate image, when the similarity is greater than a set threshold, the corresponding candidate image is used as the first candidate reference image.

[0052] At least one second candidate reference image is determined from the candidate images according to the second image feature and the first image feature associated with each target image. For example, according to the similarity between the second image feature and the first image feature of each target image, when the similarity is greater than a set threshold, the corresponding candidate image is used as the second candidate reference image. The number of first candidate reference images and second reference images can be set or adjusted according to specific needs.

[0053] Determining at least one candidate reference image based on at least one first candidate reference image and at least one second candidate reference image may include, for example, performing deduplication processing on the determined first candidate reference image and the second candidate reference image to filter out identical images in the first candidate reference image and the second candidate reference image. The sum of the number of the first candidate reference images and the second candidate reference images is greater than or equal to the number of the candidate reference images.

[0054] According to the object identity feature indicated by at least one candidate reference image, the real name information for the M images is determined. The candidate reference image includes, for example, the object identity feature, and the object identity information of the candidate reference image can be determined according to the object identity feature. According to the object identity feature of the candidate reference image, the object identity information of the M images can be determined to realize the real name of the M images.

[0055] Exemplarily, the real name information for the M images is determined based on the object identity feature indicated by at least one candidate reference image. In an exemplary manner, the M images are clustered to obtain at least one image category cluster. A second average image feature associated with each image category cluster in at least one image category cluster is determined. A target reference image in at least one candidate reference image is determined based on the first average image feature, the first image feature associated with each target image, and the second average image feature associated with each image category cluster. The real name information for the M images is determined based on the object identity feature indicated by the target reference image.

[0056] According to an embodiment of the present disclosure, at least one image category cluster is obtained by clustering M images, and a target reference image is determined from multiple candidate reference images at least based on the second average image feature of each image category cluster. The M images are clustered, and the second average image feature of each image category cluster can more accurately reflect the object identity features of the M images, and can reduce or eliminate the impact of the first average image feature offset on the result of real-name conversion.

[0057] Exemplarily, clustering is performed on M images, and at least one image category cluster is obtained according to the image features of each image in the M images. Each image category cluster may include, for example, an image set of images whose image features of a portion of the M images meet a set threshold.

[0058] For each image category cluster, a second average image feature of each image category cluster is determined according to image features of images in each image category cluster. The second average image feature may indicate the centroid of the corresponding image category cluster, for example.

[0059] A target reference image is determined from the at least one candidate reference image, for example, based on a first average image feature, a first image feature associated with each target image, and a second average image feature associated with each image category cluster.

[0060] Exemplarily, at least one determined candidate reference image is filtered according to the first average image feature, the second image feature associated with the candidate reference image, the second average image feature associated with each image category cluster, and the first image feature associated with each target image. For example, the reference images that do not meet the set conditions in the candidate reference image are screened out, and the images in the candidate reference image that are not similar to the image features of the M images are reduced, which is conducive to improving the accuracy of the real name of the M images.

[0061] The at least one reference image after filtering is sorted, for example, according to a set condition, all reference images after filtering are quality evaluated to obtain a quality evaluation value of each reference image. Each reference image is filtered according to the quality evaluation value, and a reference image that meets the preset condition is determined as a target reference image. The target reference image may include a single image or multiple images having the same object identity feature.

[0062] Through the embodiments of the present disclosure, at least one candidate reference image is determined based on the first average image feature and the first image feature associated with each target image, which can effectively improve the number and accuracy of determining the candidate reference images, and reduce the problem of inaccurate determination of the candidate reference images caused by the offset of the first average image feature when selecting the candidate reference images. By clustering M images and determining the second average image feature of each image category cluster, and determining the target reference image from multiple candidate reference images based on the second average image feature, the influence of the offset of the first average image feature on the real-name result when determining the target reference image can be effectively eliminated.

[0063] Determining at least one candidate reference image based on the first image feature associated with each target image can effectively improve the accuracy of the reference image obtained, which is beneficial to ensuring the accuracy of image real-name registration and effectively improving the efficiency of image real-name registration.

[0064] Figure 3 The flowchart of the method for image real-name conversion according to another embodiment of the present disclosure is schematically shown.

[0065] like Figure 3 As shown, in the process of image real-name conversion, process 300 of determining a target reference image in at least one candidate reference image based on a first average image feature, a first image feature associated with each target image, and a second average image feature associated with each image category cluster may include operations S310 to S330.

[0066] In operation S310, at least one candidate reference image is filtered based on the first average image feature, the second image feature associated with the candidate reference image, the second average image feature associated with each image category cluster, and the first image feature associated with each target image.

[0067] In operation S320, a quality assessment value for the filtered at least one candidate reference image is determined based on the second average image feature and the second image feature associated with the candidate reference image.

[0068] In operation S330, according to the quality evaluation value of the at least one filtered candidate reference image, a candidate reference image having a maximum quality evaluation value is used as a target reference image.

[0069] The following is an example flow of each operation of the image real-name conversion method of this embodiment.

[0070] Exemplarily, filtering at least one candidate reference image may include determining similarities between an image category cluster, a candidate reference image, and M images based on a first image feature, a second image feature, a first average image feature, and a second average image feature. For example, determining a first similarity between an image category cluster and a candidate reference image, determining a second similarity between each of the M images and the candidate reference image, and determining a third similarity between a first average image feature of the M images and the candidate reference image. The candidate reference image is filtered based on the determined first similarity, second similarity, and third similarity.

[0071] In an exemplary manner, a first similarity between each image category cluster and the candidate reference image is determined based on a second image feature associated with the candidate reference image and a second average image feature associated with each image category cluster. A second similarity between each of the M images and the candidate reference image is determined based on the first image feature and the second image feature associated with the candidate reference image. A third similarity between the first average image feature and the candidate reference image is determined based on the first average image feature and the second image feature associated with the candidate reference image. At least one candidate reference image is filtered based on the first similarity, the second similarity, and the third similarity.

[0072] According to an embodiment of the present disclosure, the candidate reference images are filtered according to the first similarity, the second similarity, and the third similarity to screen out low-correlation images in the candidate reference images. The similarity between the image features of the low-correlation images and at least one type of image features among the first image features, the first average image features, and the second average image features is lower than a preset threshold. By effectively eliminating the interference of low-correlation images, it is helpful to improve the efficiency of image real-name registration and effectively improve the effect of image real-name registration.

[0073] Exemplarily, after clustering the M images, based on the multiple image category clusters obtained by clustering, the first similarity is determined according to the second image feature and the second average image feature. The first similarity can be, for example, the number of candidate images and image category clusters whose similarities are higher than a set threshold. For example, after clustering the M images, there are 5 category clusters, and the number of candidate images with a similarity higher than 80% to the 5 categories is 3, and the candidate reference images that do not meet the threshold are filtered out.

[0074] Each of the M images has a first image feature, and a second similarity is determined based on the first image feature and a second image feature associated with the candidate image. The second similarity may include, for example, the number of images in the M images whose similarity with each candidate image is higher than a set threshold.

[0075] The third similarity is determined based on the first average image feature and the second image feature associated with the candidate reference image. The third similarity may, for example, include a preset threshold that the second image feature of the candidate image satisfies the first average image feature.

[0076] After filtering the at least one candidate reference image, a quality assessment value for the filtered at least one candidate reference image is determined based on the second average image feature and the second image feature associated with the candidate reference image.

[0077] Exemplarily, a quality assessment value weight for each image category cluster is determined based on the second average image feature. A quality assessment value of at least one filtered candidate reference image is determined based on the first similarity and the quality assessment value weight.

[0078] Each image category cluster may, for example, consist of multiple images, each image having an image quality assessment value. For any image category cluster in at least one image category cluster, a cluster quality assessment value for the image category cluster is determined based on the quality assessment value associated with each image in the image category cluster. For example, for each image in each category cluster, the quality assessment value of the image associated with each image in the image category cluster is obtained, and the quality assessment value of the maximum image is used as the cluster quality assessment value of the image category cluster.

[0079] According to the cluster quality evaluation value, the quality evaluation value weight for the image category cluster is determined. According to the cluster quality evaluation value, the quality evaluation value weight of each image category cluster is obtained by a pre-set method. For example, the quality evaluation value weight of each image category cluster can be obtained by dividing the square of the cluster quality evaluation value of each image category cluster by the sum of the squares of the quality evaluation values ​​of all image category clusters.

[0080] Exemplarily, the quality assessment value of the at least one filtered candidate reference image is determined according to the first similarity and the quality assessment value weight.

[0081] The first similarity may be determined, for example, based on a second image feature associated with the candidate reference image and a second average image feature associated with each image category cluster.

[0082] According to the determined first similarity and the quality assessment value weight, the product between the first similarity and the second average image feature of each image category cluster of the M images is calculated to obtain the quality assessment value of each filtered candidate reference image.

[0083] According to the quality evaluation value of at least one of the filtered candidate reference images, the candidate reference image with the maximum quality evaluation value is used as the target reference image. The candidate reference image corresponding to the maximum quality evaluation value has the smallest deviation from the object identity features of the M images.

[0084] According to an embodiment of the present disclosure, by determining the cluster quality assessment value of the image category cluster and determining the quality assessment weight of the image category cluster based on the cluster quality assessment value of the image category cluster, the influence of the image category cluster with a high quality assessment value on the quality assessment value of the candidate reference image can be improved, making the determined target reference image more accurate.

[0085] Figure 4 The figure schematically shows a process of image real-name conversion according to an embodiment of the present disclosure.

[0086] like Figure 4 As shown, for M images 401 to be real-named, a first average image feature 402 associated with the M images 401 is determined. For N target images 403 among the M images, a first image feature 404 associated with each of the N target images 403 is determined, and the quality assessment value of the target image is greater than a preset threshold. Based on the first average image feature 402 and the first image feature 404 associated with each target image, at least one candidate reference image 405 is determined. Based on the object identity feature 406 indicated by the at least one candidate reference image 405, the real-name information for the M images is determined.

[0087] The M images 401 are clustered to obtain at least one image category cluster 407. A second average image feature 408 associated with each image category cluster in the at least one image category cluster is determined. A target reference image 409 in at least one candidate reference image 405 is determined based on the first average image feature 402, the first image feature 404 associated with each target image, and the second average image feature 408 associated with each image category cluster. Real name information for the M images 401 is determined based on the object identity feature 406 indicated by the target reference image 409.

[0088] Through the embodiments of the present disclosure, M images are clustered to determine image category clusters, and a target reference image is determined based on the first average image feature, the first image feature associated with each target image, and the second average image feature associated with the image category cluster. This can effectively improve the accuracy of the target reference image determined from the candidate reference images and reduce the impact of the offset of the first average image feature.

[0089] Figure 5 The block diagram of an image real-name conversion device according to an embodiment of the present disclosure is schematically shown.

[0090] like Figure 5 As shown, the image real-name conversion 500 of the embodiment of the present disclosure includes, for example, a first determination module 501 , a second determination module 502 , a third determination module 503 and a fourth determination module 504 .

[0091] The first determination module 501 is configured to determine, for M images to be real-named, a first average image feature associated with the M images, where M is an integer greater than 1; the second determination module 502 is configured to determine, for N target images among the M images, a first image feature associated with each of the N target images, where a quality assessment value of the target image is greater than a preset threshold, where N is an integer greater than 1 and less than M; the third determination module 503 is configured to determine at least one candidate reference image based on the first average image feature and the first image feature associated with each target image; the fourth determination module 504 is configured to determine the real-name information for the M images based on the object identity feature indicated by the at least one candidate reference image.

[0092] Through the embodiments of the present disclosure, at least one candidate reference image is determined based on the first average image feature and the first image feature associated with each target image, which can effectively improve the accuracy of determining the candidate reference image and reduce the problem of inaccurate determination of the candidate reference image caused by the offset of the first average image feature. By determining at least one candidate reference image based on the first image feature associated with each target image, the accuracy of the acquired reference image can be effectively improved, which is conducive to more accurate real-name registration of M images.

[0093] According to an embodiment of the present disclosure, the fourth determination module includes: a first determination submodule, configured to cluster M images to obtain at least one image category cluster; a second determination submodule, configured to determine a second average image feature associated with each image category cluster in at least one image category cluster; a third determination submodule, configured to determine a target reference image in at least one candidate reference image based on the first average image feature, the first image feature associated with each target image, and the second average image feature associated with each image category cluster; and a fourth determination submodule, configured to determine real-name information for the M images based on an object identity feature indicated by the target reference image.

[0094] According to an embodiment of the present disclosure, the third determination module includes: a first acquisition module, configured to acquire a second image feature for each image in the candidate images; a fifth determination submodule, configured to determine at least one first candidate reference image from the candidate images based on the first average image feature and the second image feature; a sixth determination submodule, configured to determine at least one second candidate reference image from the candidate images based on the second image feature and the first image feature associated with each target image; and a seventh determination submodule, configured to determine at least one candidate reference image based on the at least one first candidate reference image and the at least one second candidate reference image.

[0095] According to an embodiment of the present disclosure, the third determination submodule includes: a processing unit, configured to filter at least one candidate reference image based on a first average image feature, a second image feature associated with the candidate reference image, a second average image feature associated with each image category cluster, and a first image feature associated with each target image; a determination unit, configured to determine a quality assessment value for at least one filtered candidate reference image based on the second average image feature and the second image feature associated with the candidate reference image; and an assessment unit, configured to select the candidate reference image with the maximum quality assessment value as the target reference image based on the quality assessment value for at least one filtered candidate reference image.

[0096] According to an embodiment of the present disclosure, a processing unit includes a processing subunit, which is configured to: determine a first similarity between each image category cluster and the candidate reference image based on a second image feature associated with the candidate reference image and a second average image feature associated with each image category cluster; determine a second similarity between each of M images and the candidate reference image based on the first image feature and the second image feature associated with the candidate reference image; determine a third similarity between the first average image feature and the candidate reference image based on the first average image feature and the second image feature associated with the candidate reference image; and filter at least one candidate reference image based on the first similarity, the second similarity and the third similarity.

[0097] According to an embodiment of the present disclosure, the determination unit includes: a first determination subunit, configured to determine a quality assessment value weight for each image category cluster based on a second average image feature; and a second determination subunit, configured to determine a quality assessment value of at least one filtered candidate reference image based on a first similarity and a quality assessment value weight.

[0098] According to an embodiment of the present disclosure, the first determination subunit includes a weight determination subunit, and the weight determination subunit is configured to: for any image category cluster in at least one image category cluster, determine a cluster quality assessment value for the image category cluster based on a quality assessment value associated with each image in the image category cluster; and determine a quality assessment value weight for the image category cluster based on the cluster quality assessment value.

[0099] It should be noted that the information collection, storage, use, processing, transmission, provision and disclosure involved in the technical solution of the present disclosure are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0100] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0101] Figure 6A block diagram of an electronic device for executing an image real-name conversion method according to an embodiment of the present disclosure is schematically shown.

[0102] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device 600 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0103] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0104] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0105] The computing unit 601 may include various general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as the image real-name method. For example, in some embodiments, the image real-name method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the image real-name method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the image real-name method in any other appropriate manner (e.g., by means of firmware).

[0106] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0107] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable path planning device, so that the program code, when executed by the processor or controller, implements the functions / operations specified in the flow chart and / or block diagram. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0108] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0109] To provide interaction with an object, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the object; and a keyboard and pointing device (e.g., a mouse or trackball) through which the object can provide input to the computer. Other types of devices can also be used to provide interaction with an object; for example, the feedback provided to the object can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the object can be received in any form (including acoustic input, voice input, or tactile input).

[0110] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., an object computer having a graphical object interface or a web browser through which the object can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communications network). Examples of communications networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0111] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0112] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0113] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for real-name registration of an image, comprising: For M images to be annotated, determining a first average image feature associated with the M images, where M is an integer greater than 1; For N target images among the M images, determining a first image feature associated with each of the N target images, wherein a quality assessment value of the target image is greater than a preset threshold, and N is an integer greater than 1 and less than M; Determining at least one candidate reference image according to the first average image feature and the first image feature associated with each target image, comprising: acquiring a second image feature for each image in the candidate images; determining at least one first candidate reference image from the candidate images according to the first average image feature and the second image feature; determining at least one second candidate reference image from the candidate images according to the second image feature and the first image feature associated with each target image; and determining at least one candidate reference image according to the at least one first candidate reference image and the at least one second candidate reference image; and Real name information for the M images is determined based on the object identity feature indicated by the at least one candidate reference image.

2. The method according to claim 1, wherein: The determining the real name information for the M images according to the object identity feature indicated by the at least one candidate reference image comprises: Clustering the M images to obtain at least one image category cluster; determining a second average image feature associated with each of the at least one image category cluster; determining a target reference image among the at least one candidate reference image based on the first average image feature, the first image feature associated with each target image, and the second average image feature associated with each image category cluster; and The real name information for the M images is determined according to the object identity feature indicated by the target reference image.

3. The method according to claim 1, wherein: The step of determining a target reference image in the at least one candidate reference image according to the first average image feature, the first image feature associated with each target image, and the second average image feature associated with each image category cluster comprises: filtering the at least one candidate reference image based on the first average image feature, the second image feature associated with the candidate reference image, the second average image feature associated with each of the image category clusters, and the first image feature associated with each of the target images; determining a quality assessment value for at least one filtered candidate reference image based on the second average image feature and a second image feature associated with the candidate reference image; and According to the quality evaluation value of at least one filtered candidate reference image, a candidate reference image with a maximum quality evaluation value is used as a target reference image.

4. The method according to claim 3, wherein: The filtering of the at least one candidate reference image according to the first average image feature, the second image feature associated with the candidate reference image, the second average image feature associated with each image category cluster, and the first image feature associated with each target image comprises: determining a first similarity between each image category cluster and the candidate reference image according to a second image feature associated with the candidate reference image and a second average image feature associated with each image category cluster; determining a second similarity between each of the M images and the candidate reference image based on the first image feature and the second image feature associated with the candidate reference image; determining a third similarity between the first average image feature and the candidate reference image based on the first average image feature and a second image feature associated with the candidate reference image; and The at least one candidate reference image is filtered according to the first similarity, the second similarity and the third similarity.

5. The method according to claim 4, wherein: The step of determining a quality assessment value for at least one filtered candidate reference image according to the second average image feature and a second image feature associated with the candidate reference image comprises: Determining a quality assessment value weight for each image category cluster according to the second average image feature; and A quality assessment value of at least one filtered candidate reference image is determined according to the first similarity and the quality assessment value weight.

6. The method according to claim 5, wherein: Determining the quality assessment value weight for each image category cluster according to the second average image feature includes: For any image category cluster in the at least one image category cluster, determining a cluster quality assessment value for the image category cluster according to the quality assessment value associated with each image in the image category cluster; and A quality assessment value weight for the image category cluster is determined according to the cluster quality assessment value.

7. An image real-name device, comprising: A first determination module is configured to determine, for M images to be annotated, a first average image feature associated with the M images, where M is an integer greater than 1; a second determination module configured to determine, for N target images among the M images, a first image feature associated with each of the N target images, wherein a quality assessment value of the target image is greater than a preset threshold, and N is an integer greater than 1 and less than M; a third determination module, configured to determine at least one candidate reference image according to the first average image feature and a first image feature associated with each of the target images; and a fourth determination module configured to determine the real name information for the M images according to the object identity feature indicated by the at least one candidate reference image; Wherein, the third determination module includes: A first acquisition module is configured to acquire a second image feature for each image in the candidate images; a fifth determination submodule, configured to determine at least one first candidate reference image from the candidate images according to the first average image feature and the second image feature; a sixth determination submodule, configured to determine at least one second candidate reference image from the candidate images according to the second image feature and the first image feature associated with each target image; and The seventh determination submodule is configured to determine at least one candidate reference image according to the at least one first candidate reference image and the at least one second candidate reference image.

8. The device according to claim 7, wherein: The fourth determination module comprises: A first determination submodule is configured to cluster the M images to obtain at least one image category cluster; a second determination submodule configured to determine a second average image feature associated with each of the at least one image category cluster; a third determination submodule, configured to determine a target reference image among the at least one candidate reference image according to the first average image feature, the first image feature associated with each target image, and the second average image feature associated with each image category cluster; and The fourth determination submodule is configured to determine the real name information for the M images according to the object identity feature indicated by the target reference image.

9. The device according to claim 7, wherein: The third determining submodule comprises: a processing unit configured to filter the at least one candidate reference image based on the first average image feature, the second image feature associated with the candidate reference image, the second average image feature associated with each of the image category clusters, and the first image feature associated with each of the target images; a determining unit configured to determine a quality assessment value for at least one filtered candidate reference image based on the second average image feature and a second image feature associated with the candidate reference image; and The evaluation unit is configured to select a candidate reference image having a maximum quality evaluation value as a target reference image according to the quality evaluation value of at least one candidate reference image after filtering.

10. The device according to claim 9, wherein: The processing unit includes a processing subunit, and the processing subunit is configured as follows: determining a first similarity between each image category cluster and the candidate reference image according to a second image feature associated with the candidate reference image and a second average image feature associated with each image category cluster; determining a second similarity between each of the M images and the candidate reference image based on the first image feature and the second image feature associated with the candidate reference image; determining a third similarity between the first average image feature and the candidate reference image based on the first average image feature and a second image feature associated with the candidate reference image; as well as The at least one candidate reference image is filtered according to the first similarity, the second similarity and the third similarity.

11. The device according to claim 10, wherein: The determining unit comprises: A first determining subunit is configured to determine a quality assessment value weight for each image category cluster according to the second average image feature; The second determining subunit is configured to determine the quality assessment value of at least one filtered candidate reference image according to the first similarity and the quality assessment value weight.

12. The device according to claim 11, wherein The first determination subunit includes a weight determination subunit, The weight determination subunit is configured as follows: For any image category cluster in the at least one image category cluster, determining a cluster quality assessment value for the image category cluster according to a quality assessment value associated with each image in the image category cluster; as well as A quality assessment value weight for the image category cluster is determined according to the cluster quality assessment value.

13. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 6.

15. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Citation Information

Patent Citations

  • Archiving method and device

    CN109740003A

  • Data processing method and related product

    CN109815823A

  • Method and device for determining identity type of person and electronic system

    CN111666441A

  • Image data processing method and device, computer equipment and storage medium

    CN112597984A

  • Image recognition method and device, electronic equipment and medium

    CN113344055A