Person image clustering method and device, computer device and storage medium

Abstract

Embodiments of the present application relate to a person image clustering method and device, computer equipment and a storage medium, the method comprising: obtaining face feature information corresponding to each person image in a person image set; determining face feature similarity between each person image and other person images based on the face feature information; clustering person images with face feature similarity greater than a first threshold value to obtain a plurality of first person image sets; performing optimized classification on person images in each first person image set to obtain a plurality of second person image sets; comparing set similarity between the plurality of second person image sets, and clustering second person image sets with set similarity greater than a second threshold value to obtain a third person image set. By this method, the problem of different person images under the same person identity being clustered in different sets can be avoided, and accurate clustering of person images can be achieved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of human image processing, and in particular to a human image clustering method and device, computer equipment and a storage medium. BACKGROUND

[0002] The current human image clustering algorithm is based on the strategy of comparing human face or body features, which compares the human face or body feature set uploaded by the device with the database basic feature set, associates the target feature with the basic feature set relationship through the comparison of similarity and the corresponding preset similarity threshold, and completes the human image clustering task.

[0003] However, the feature discrimination of human images under different light, angle, and occlusion conditions may be different, so human images under different collection conditions may have a multi-domain problem, that is, different human images under the same human identity will be clustered in different sets. How to accurately cluster different human images under the same human identity under different collection conditions has become a problem to be solved. SUMMARY

[0004] In view of this, in order to solve the above or part of the technical problems, embodiments of the present application provide a human image clustering method, device, computer equipment and storage medium.

[0005] In a first aspect, embodiments of the present application provide a human image clustering method, comprising:

[0006] Obtaining the human face feature information corresponding to each human image in the human image set;

[0007] Determining the human face feature similarity between each human image and other human images based on the human face feature information;

[0008] Clustering the human images with a human face feature similarity greater than a first threshold value to obtain a plurality of first human image sets;

[0009] Optimizing the classification of human images in each first human image set to obtain a plurality of second human image sets;

[0010] Comparing the set similarity between the plurality of second human image sets, and clustering the second human image sets with a set similarity greater than a second threshold value to obtain a third human image set.

[0011] In a second aspect, embodiments of the present application provide a human image clustering device, comprising:

[0012] The acquisition module is configured to obtain the human face feature information corresponding to each human image in the human image set;

[0013] determining module, configured to determine face feature similarity between each of the person images and other person images based on the face feature information;

[0014] clustering module, configured to cluster the person images with face feature similarity greater than a first threshold to obtain a plurality of first person image sets;

[0015] The clustering module is further configured to perform optimized classification on the person images in each of the first person image sets to obtain a plurality of second person image sets.

[0016] The clustering module is further configured to compare set similarity between the plurality of second person image sets, and cluster the second person image sets with set similarity greater than a second threshold to obtain a third person image set.

[0017] In a third aspect, an embodiment of the present application provides a computer device, comprising a processor and a memory, wherein the processor is configured to execute a person image clustering program stored in the memory to implement the person image clustering method in the first aspect.

[0018] In a fourth aspect, an embodiment of the present application provides a storage medium, comprising one or more programs stored in the storage medium, wherein the one or more programs can be executed by one or more processors to implement the person image clustering method in the first aspect.

[0019] The person image clustering scheme provided by the embodiment of the present application comprises the following steps: obtaining face feature information corresponding to each of the person images in a person image set; determining face feature similarity between each of the person images and other person images based on the face feature information; clustering the person images with face feature similarity greater than a first threshold to obtain a plurality of first person image sets; performing optimized classification on the person images in each of the first person image sets to obtain a plurality of second person image sets; comparing set similarity between the plurality of second person image sets, and clustering the second person image sets with set similarity greater than a second threshold to obtain a third person image set. Compared with the prior art, the present scheme can avoid the problem that different person images under the same person identity are clustered in different sets, and can realize accurate clustering of person images. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 A flowchart of a person image clustering method provided by the embodiment of the present application is shown in the figure;

[0021] Figure 2 Another flowchart of a person image clustering method provided by an embodiment of the present application is shown in FIG. 6.

[0022] Figure 3 A flowchart of a method for obtaining set similarity between second person image sets provided by an embodiment of the present application is shown in FIG. 7.

[0023] Figure 4 A flowchart of another method for obtaining set similarity between second person image sets provided by an embodiment of the present application is shown in FIG. 8.

[0024] Figure 5 A structural diagram of a person image clustering device provided by an embodiment of the present application is shown in FIG. 9.

[0025] Figure 6 A structural diagram of a computer device provided by an embodiment of the present application is shown in FIG. 10. DETAILED DESCRIPTION

[0026] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0027] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0028] Figure 1 A flowchart of a person image clustering method provided by an embodiment of the present application is shown in FIG. 6. Figure 1 The method specifically includes:

[0029] S11, obtaining face feature information corresponding to each person image in a person image set.

[0030] The present application can be applied to a person image clustering scene and can be used to identify persons in different backgrounds and perform image clustering under the same person identity. In the embodiments of the present application, a person image set containing multiple person images is given first, and then a face detector and a face deep feature recognition model are used to identify the face feature information corresponding to each person image.

[0031] Specifically, the face detector can be a face recognition model, and the image of the person to be detected is input into the face recognition model, the face recognition model recognizes the received image of the person, and the face position in the image of the person is circled, and a face frame can be used to identify the recognized face region.

[0032] Further, the face feature information of the face in the face frame in each image of the person is recognized by a face depth feature recognition model, the face depth feature recognition model stores standard face feature information, the face feature information corresponding to the recognized image of the person is compared with the standard face feature information, and the face feature information corresponding to each image of the person in the image set is obtained, which can be represented as a feature vector.

[0033] S12, determining the face feature similarity between each image of the person and other images of the person based on the face feature information.

[0034] Based on the face feature information corresponding to each image of the person in the image set obtained above, the face feature similarity between each image of the person and other images of the person in the image set can be compared, and the face feature similarity can be a cosine similarity.

[0035] S13, clustering the images of the person with the face feature similarity greater than a first threshold value to obtain a plurality of first image sets of the person.

[0036] In the embodiment of the application, a first threshold value, i.e., a face feature similarity threshold value, is preset, and the images of the person with the face feature similarity greater than the first threshold value are clustered to obtain a plurality of first image sets of the person.

[0037] For example, the image set of the person contains five images of the person, the face feature similarity between image A and image B is 9, the face feature similarity between image C and image D is 9, the face feature similarity between image A and image E is 1, and the face feature similarity between image C and image E is 2; the first threshold value is set to 8, then image A and image B can be clustered into the same set, image C and image D can be clustered into the same set, and image E can be a separate set.

[0038] S14, optimizing the classification of the images of the person in each first image set of the person to obtain a plurality of second image sets of the person.

[0039] In the embodiment of the present application, the obtained multiple first person image sets are classified again, and the optimization standard can be to refine multiple feature categories, and the images in each first person image set that meet the feature categories are clustered again, and each first person image set can obtain multiple second person image sets. Wherein, the features of images of different angles, images under different illuminations and the like can be set as feature categories, and how to classify again will be described in detail in the following embodiments, and will not be described in detail here.

[0040] S15, comparing the set similarity between the multiple second person image sets, clustering the second person image sets with set similarity greater than the second threshold value to obtain third person image sets.

[0041] In the embodiment of the present application, a second threshold value, i.e. a set similarity threshold value, can be set in advance, and after obtaining multiple second person image sets, the set similarity between the multiple second person image sets can be compared, and the second person image sets with set similarity greater than the second threshold value can be clustered to obtain third person image sets, which are the final clustering results and contain all images under the same person identity. How to compare the set similarity will be described in detail in the following embodiments, and will not be described in detail here.

[0042] The person image clustering method provided by the embodiment of the present application comprises the following steps:

[0043] Figure 2 Another flowchart of the person image clustering method provided by the embodiment of the present application is shown in FIG. 2, and the method specifically comprises the following steps: Figure 2

[0044] S21, identifying the face picture corresponding to each person image in the person image set.​

[0045] In the embodiment of the present application, first, a person image set containing multiple person images is given, and then a face detector and a face deep feature recognition model are used to recognize the face feature information corresponding to each person image.

[0046] Specifically, the face detector recognizes and encircles the face position in the person image, which can be marked by a face recognition frame to obtain the face picture corresponding to each person image.

[0047] S22, face feature information is extracted from the face picture to obtain the face feature information corresponding to each person image in the person image set.

[0048] The face feature information of each person image corresponding face picture is recognized by the face deep feature recognition model to obtain the face feature information corresponding to each person image in the person image set.

[0049] S23, based on the face feature information, the face feature similarity between each person image and other person images is determined.

[0050] Based on the face feature information corresponding to each person image in the person image set obtained above, the face feature similarity between each person image and other person images in the person image set can be compared. The face feature similarity can be cosine similarity.

[0051] For example, the person image set contains 4 person images, the face feature similarity between image A and image B is 9, the face feature similarity between image A and image C is 2, the face feature similarity between image A and image D is 4, the face feature similarity between image B and image C is 3, the face feature similarity between image B and image D is 3, and the face feature similarity between image C and image D is 8.

[0052] S24, based on the face feature similarity between each person image and other person images, a similarity network structure diagram is constructed.

[0053] Based on the face feature similarity between each person image and other person images obtained above, a similarity network structure diagram is constructed, wherein each node in the similarity network structure diagram is a face picture, and the connection line between each two nodes is the face feature similarity.

[0054] S25, the connection line between each two nodes in the similarity network structure diagram is reserved if the face feature similarity corresponding to the connection line is greater than a first threshold value, and the remaining connection lines are deleted to obtain multiple groups of person images with connection relationship.

[0055] S26, cluster the images of the persons in each group having the connection relationship to obtain a plurality of first image sets of persons.

[0056] The following uniformly describes S25-S26:

[0057] In the embodiment of the application, a first threshold, i.e., a face feature similarity threshold, can be set, and the connection lines between each two nodes in the similarity network structure graph are retained if the face feature similarity corresponding to the connection lines is greater than the first threshold, and the connection lines are deleted if the face feature similarity is less than or equal to the first threshold, thereby obtaining a plurality of groups of images of persons having the connection relationship.

[0058] Further, each group of images of persons having the connection relationship shown in the similarity network structure graph is taken as a clustering set, thereby obtaining a plurality of first image sets of persons.

[0059] S27, determine a plurality of feature categories based on the face feature information corresponding to each image of persons in the first image set of persons.

[0060] S28, perform an optimized classification on the images of persons in the first image set of persons based on the feature categories, thereby obtaining a plurality of second image sets of persons.

[0061] The following uniformly describes S27-S28:

[0062] In the embodiment of the application, the plurality of first image sets of persons obtained above are subjected to a secondary optimized classification, and the optimization standard can be to refine the plurality of feature categories, and the images in each first image set of persons that meet the feature categories are clustered again, thereby obtaining a plurality of second image sets of persons. The features of images at different angles, images under different illuminations, etc. can be set as the feature categories.

[0063] S29, compare the set similarities between the plurality of second image sets of persons, and cluster the second image sets of persons having the set similarity greater than a second threshold, thereby obtaining a third image set of persons.

[0064] In the embodiment of the application, a second threshold, i.e., a set similarity threshold, can be set in advance, and after the plurality of second image sets of persons are obtained, the set similarities between the plurality of second image sets of persons can be compared, and the second image sets of persons having the set similarity greater than the second threshold are clustered, thereby obtaining a third image set of persons, which is taken as a final clustering result and contains all images under the same identity of persons.

[0065] Specifically, the set similarity can include any one of an average feature similarity, a maximum feature similarity, a minimum feature similarity, and a central feature similarity. Taking the average feature similarity as an example, the average feature similarity is an average value of face feature similarities between any two person images in two second person image sets.

[0066] For example, set 1 includes three person images A, B, and C, and set 2 includes three person images a, b, and c. The similarity between A and a is 4, the similarity between A and b is 8, the similarity between A and c is 6, the similarity between B and a is 5, the similarity between B and b is 4, the similarity between B and c is 9, the similarity between C and a is 8, the similarity between C and b is 2, and the similarity between C and c is 7. The average feature similarity between set 1 and set 2 is (4+8+6+5+4+9+8+2+7) ÷ 9 = 5.89, and the average feature similarity threshold (the second threshold) is set to 5. It can be determined that the average feature similarity between set 1 and set 2 is greater than the average feature similarity threshold, which represents that set 1 and set 2 are image sets under the same person identity, and set 1 and set 2 are clustered.

[0067] The person image clustering method provided by the embodiment of the present application comprises the following steps: obtaining face feature information corresponding to each person image in a person image set; determining face feature similarities between each person image and other person images based on the face feature information; clustering person images with a face feature similarity greater than a first threshold to obtain a plurality of first person image sets; performing optimized classification on person images in each first person image set to obtain a plurality of second person image sets; comparing set similarities between the plurality of second person image sets, and clustering second person image sets with a set similarity greater than a second threshold to obtain third person image sets. By multiple clustering and similarity comparison between different clustering sets, the method can avoid the problem that different person images under the same person identity are clustered in different sets, and realizes accurate clustering of person images.

[0068] Figure 3 A flowchart of a method for obtaining set similarities between second person image sets provided by the embodiment of the present application is shown in FIG. 1. Figure 3 The method specifically comprises the following steps.

[0069] S31, obtaining face feature similarities corresponding to any two person images in any two second person image sets.

[0070] S32, determining set similarities between the plurality of second person image sets based on the face feature similarities, wherein the set similarity is any one of an average feature similarity, a maximum feature similarity, and a minimum feature similarity.

[0071] The following uniformly describes S31-S32:

[0072] In the embodiment of the application, the face feature similarity of any two person images in any two second person image sets is obtained, and the set similarity is calculated according to the face feature similarity. The set similarity can include any one of the average feature similarity, the maximum feature similarity, the minimum feature similarity and the central feature similarity. Taking the maximum feature similarity as an example, the maximum feature similarity is the maximum value of the face feature similarity of any two person images between the two second person image sets.

[0073] Specifically, for example, the set 1 includes three person images A, B and C, and the set 2 includes three person images a, b and c. The similarity of A and a is 4, the similarity of A and b is 8, the similarity of A and c is 6, the similarity of B and a is 5, the similarity of B and b is 4, the similarity of B and c is 9, the similarity of C and a is 8, the similarity of C and b is 2, and the similarity of C and c is 7. The maximum feature similarity of the set 1 and the set 2 is 9, and the maximum feature similarity threshold (the second threshold) is set to 8. It can be determined that the average feature similarity of the set 1 and the set 2 is greater than the maximum feature similarity threshold, which represents that the set 1 and the set 2 are image sets under the same person identity, and the set 1 and the set 2 are clustered.

[0074] Correspondingly, the minimum feature similarity is the minimum value of the face feature similarity of any two person images between the two second person image sets. If the minimum feature similarity is less than or equal to the minimum feature similarity threshold, it can be determined that the two sets are not image sets under the same person identity. If the minimum feature similarity is greater than the minimum feature similarity threshold, it can be determined that the two sets are image sets under the same person identity, and the two sets are clustered.

[0075] The method for obtaining the set similarity between the second person image sets provided in the embodiment of the application can perform secondary clustering on the second person image sets, can avoid the problem that different person images under the same person identity are clustered in different sets, and can realize accurate clustering of person images.

[0076] Figure 4 Another flowchart of the method for obtaining the set similarity between the second person image sets provided in the embodiment of the application is shown in FIG. 4B. Figure 4 The method specifically includes the following steps.

[0077] S41, obtaining a central person image corresponding to each second person image set.

[0078] In the embodiment of the present application, the betweenness centrality is used to measure the central node of each second person image set. The core node is the node with the highest frequency in all the shortest paths between any two nodes in the similarity network structure graph, and the person image corresponding to the core node is the central person image corresponding to the second person image set.

[0079] In S42, the central feature similarity between the second person image sets is determined based on the face feature information corresponding to the central person image.

[0080] In S43, the central feature similarity is used to determine the set similarity between the second person image sets.

[0081] The following uniformly describes S42-S43:

[0082] Based on the face feature information corresponding to the central person image, the central feature similarity between the second person image sets can be compared.

[0083] For example, the central feature similarity between set 1 and set 2 is 8, the central feature similarity between set 1 and set 3 is 2, the central feature similarity between set 1 and set 4 is 4, the central feature similarity between set 2 and set 3 is 4, the central feature similarity between set 2 and set 4 is 1, the central feature similarity between set 3 and set 4 is 9, and the central feature similarity threshold is set to 5. It can be determined that the central feature similarity between set 1 and set 2 is greater than the central feature similarity threshold, which represents that set 1 and set 2 are image sets under the same person identity, and set 1 and set 2 are clustered. The central feature similarity between set 3 and set 4 is greater than the central feature similarity threshold, which represents that set 3 and set 4 are image sets under the same person identity, and set 3 and set 4 are clustered.

[0084] The method for obtaining the set similarity between the second person image sets provided in the embodiment of the present application can perform secondary clustering on the second person image sets, which can avoid the problem that different person images under the same person identity are clustered in different sets, and realizes accurate clustering of person images.

[0085] Figure 5 The structure diagram of a person image clustering device provided in the embodiment of the present application is specifically shown in the following figure:

[0086] The obtaining module 501 is configured to obtain the face feature information corresponding to each person image in the person image set.

[0087] The determining module 502 is configured to determine the face feature similarity between each person image and other person images based on the face feature information.

[0088] The clustering module 503 is configured to cluster the person images with the face feature similarity greater than the first threshold value to obtain a plurality of first person image sets.

[0089] The clustering module 503 is further configured to perform optimized classification on the person images in each first person image set to obtain a plurality of second person image sets.

[0090] The clustering module 503 is further configured to compare set similarities between the plurality of second person image sets, and cluster the second person image sets with the set similarity greater than a second threshold value to obtain a third person image set.

[0091] In one possible implementation, the acquisition module 501 is specifically configured to identify a face picture corresponding to each person image in the person image set, and extract face feature information from the face picture to obtain face feature information corresponding to each person image in the person image set.

[0092] In one possible implementation, the clustering module 503 is specifically configured to construct a similarity network structure diagram based on the face feature similarity between each person image and other person images, wherein each node in the similarity network structure diagram is a face picture, and a connection line between each two nodes is the face feature similarity; the clustering module 503 is further configured to retain a connection line with the face feature similarity greater than a first threshold value between each two nodes in the similarity network structure diagram, and delete the remaining connection lines to obtain a plurality of groups of person images with a connection relationship; and the clustering module 503 is further configured to cluster each group of person images with the connection relationship to obtain a plurality of first person image sets.

[0093] In one possible implementation, the clustering module 503 is further configured to determine a plurality of feature categories based on the face feature information corresponding to each person image in the first person image set, and perform optimized classification on the person images in the first person image set based on the feature categories to obtain a plurality of second person image sets.

[0094] In one possible implementation, the clustering module 503 is further configured to obtain face feature similarities corresponding to any two person images in any two second person image sets, and determine a set similarity between the plurality of second person image sets based on the face feature similarities, the set similarity being any one of an average feature similarity, a maximum feature similarity, and a minimum feature similarity.

[0095] In a possible implementation, the clustering module 503 is further configured to acquire a center person image corresponding to each second person image set; determine a center feature similarity between the second person image sets based on face feature information corresponding to the center person image; and determine a set similarity between the second person image sets based on the center feature similarity.

[0096] The person image clustering apparatus provided in the embodiment can be a person image clustering apparatus as shown in Figure 5 The person image clustering apparatus can perform all steps of the person image clustering method as shown in Figures 1-4 The person image clustering apparatus can realize the technical effects of the person image clustering method as shown in Figures 1-4 The person image clustering apparatus can realize the technical effects of the person image clustering method as shown in Figures 1-4 For brevity, the related description is not repeated here.

[0097] Figure 6 A structural schematic diagram of a computer device provided in the embodiment of the present application is shown in Figure 6 The computer device 600 shown in the embodiment of the present application includes at least one processor 601, a memory 602, at least one network interface 604 and other user interfaces 603. The various components in the computer device 600 are coupled together through a bus system 605. It can be understood that the bus system 605 is used to realize the connection and communication between the components. The bus system 605 includes not only a data bus, but also a power bus, a control bus and a status signal bus. However, for the purpose of clear illustration, all kinds of buses are marked as the bus system 605 in the embodiment of the present application. Figure 6

[0098] The user interface 603 can include a display, a keyboard or a clicking device (for example, a mouse, a trackball, a touchpad or a touch screen, etc.).

[0099] ​It is to be appreciated that the memory 602 in the embodiments of the present application can be a volatile memory or a nonvolatile memory, or can include both volatile and nonvolatile memory. Among them, the nonvolatile memory can be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically EPROM (EEPROM), or a flash memory. The volatile memory can be a Random Access Memory (RAM) used as an external cache. By way of example, and not limitation, many forms of RAM can be used, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synch link DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memory 602 described herein is intended to include, without being limited to, these and any other suitable types of memory.

[0100] In some embodiments, the memory 602 stores the following elements, executable units or data structures, or a subset of them, or an extended set of them: an operating system 6021 and application programs 6022.

[0101] Among them, the operating system 6021 includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application programs 6022 include various application programs, such as a media player (Media Player), a browser (Browser), etc., for implementing various application services. The programs for implementing the method embodiments of the present application can be included in the application programs 6022.

[0102] In the embodiments of the present application, by calling the programs or instructions stored in the memory 602, specifically, the programs or instructions stored in the application programs 6022, the processor 601 is used to execute the method steps provided by each method embodiment, for example, including:

[0103] obtaining face feature information corresponding to each of the person images in the person image set; determining face feature similarity between each of the person images and other person images based on the face feature information; clustering person images with face feature similarity greater than a first threshold value to obtain a plurality of first person image sets; performing optimized classification on the person images in each of the first person image sets to obtain a plurality of second person image sets; and clustering second person image sets with set similarity greater than a second threshold value to obtain a third person image set.

[0104] In one possible implementation, a face picture corresponding to each of the person images in the person image set is identified; and face feature information is extracted from the face picture to obtain face feature information corresponding to each of the person images in the person image set.

[0105] In one possible implementation, a similarity network structure diagram is constructed based on the face feature similarity between each of the person images and other person images, wherein each node in the similarity network structure diagram is a face picture, and a connection line between each two nodes is the face feature similarity; connection lines with face feature similarity greater than a first threshold value between each two nodes in the similarity network structure diagram are retained, and the remaining connection lines are deleted to obtain a plurality of groups of person images with a connection relationship; and each group of person images with a connection relationship is clustered to obtain a plurality of first person image sets.

[0106] In one possible implementation, a plurality of feature categories are determined based on the face feature information corresponding to each of the person images in the first person image set; and the person images in the first person image set are optimized classified based on the feature categories to obtain a plurality of second person image sets.

[0107] In one possible implementation, the set similarity between the plurality of second person image sets is obtained by: obtaining face feature similarity corresponding to any two person images in any two of the second person image sets; and determining the set similarity between the plurality of second person image sets based on the face feature similarity, the set similarity being any one of an average feature similarity, a maximum feature similarity, and a minimum feature similarity.

[0108] In one possible implementation, the set similarity between the plurality of second person image sets is obtained by: obtaining a center person image corresponding to each of the second person image sets; determining center feature similarity between the plurality of second person image sets based on face feature information corresponding to the center person image; and determining the set similarity between the plurality of second person image sets based on the center feature similarity.

[0109] The method disclosed by the embodiments of the present application can be applied to the processor 601 or implemented by the processor 601. The processor 601 can be an integrated circuit chip having a signal processing capability. In the implementation process, the steps of the method can be completed by an integrated logic circuit or an instruction in the form of software in the processor 601. The processor 601 described above can be a general processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component. The disclosed methods, steps and logic block diagrams in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as a hardware code processor for execution, or a combination of hardware and software units in the code processor for execution. The software unit can be located in a random access memory, a flash memory, a read only memory, a programmable read only memory or an electrically erasable programmable memory, a register or other mature storage medium in the art. The storage medium is located in the memory 602, and the processor 601 reads the information in the memory 602 and combines the hardware to complete the steps of the above method.

[0110] It can be understood that the embodiments described herein can be realized by hardware, software, firmware, middleware, microcode or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), general purpose processors, controllers, micro-controllers, microprocessors, other electronic units for executing the functions described in the present application or a combination thereof.

[0111] For software implementation, the technology described herein can be implemented by units performing the functions described herein. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0112] The computer device provided by the embodiment can be a computer device as shown in Figure 6 The computer device can execute all steps of the person image clustering method as shown in Figures 1-4 The computer device can realize the technical effects of the person image clustering method as shown in Figures 1-4 For brevity, the related description is not repeated here. Figures 1-4

[0113] The embodiment of the present application further provides a storage medium (computer readable storage medium). The storage medium stores one or more programs. The storage medium can include a volatile memory such as a random access memory, and the memory can also include a non-volatile memory such as a read-only memory, a flash memory, a hard disk or a solid state disk, and the memory can also include a combination of the above kinds of memories.

[0114] When the one or more programs in the storage medium are executed by one or more processors, the person image clustering method executed on the computer device side described above can be realized.

[0115] The processor is configured to execute the person image clustering program stored in the memory, so as to realize the following steps of the person image clustering method executed on the computer device side:

[0116] Obtain the face feature information corresponding to each person image in the person image set; determine the face feature similarity between each person image and other person images based on the face feature information; cluster the person images with face feature similarity greater than a first threshold value to obtain a plurality of first person image sets; perform optimized classification on the person images in each first person image set to obtain a plurality of second person image sets; compare the set similarity between the plurality of second person image sets, and cluster the second person image sets with set similarity greater than a second threshold value to obtain a third person image set.

[0117] In one possible implementation, a face picture corresponding to each person image in the person image set is identified; face feature information is extracted from the face picture to obtain the face feature information corresponding to each person image in the person image set.

[0118] ​In a possible implementation, a similarity network structure graph is constructed based on the face feature similarity between each person image and other person images, wherein each node in the similarity network structure graph is a face picture, and a connection line between each two nodes is the face feature similarity; connection lines with a face feature similarity greater than a first threshold value between each two nodes in the similarity network structure graph are retained, and the remaining connection lines are deleted, to obtain a plurality of groups of person images with a connection relationship; each group of person images with a connection relationship is clustered to obtain a plurality of first person image sets.

[0119] In a possible implementation, a plurality of feature categories are determined based on the face feature information corresponding to each person image in the first person image set; and the person images in the first person image set are classified based on the feature categories, to obtain a plurality of second person image sets.

[0120] In a possible implementation, the set similarity between the plurality of second person image sets is obtained by: obtaining the face feature similarity corresponding to any two person images in any two second person image sets; and determining the set similarity between the plurality of second person image sets based on the face feature similarity, wherein the set similarity is any one of an average feature similarity, a maximum feature similarity, and a minimum feature similarity.

[0121] In a possible implementation, the set similarity between the plurality of second person image sets is obtained by: obtaining a center person image corresponding to each second person image set; determining a center feature similarity between the plurality of second person image sets based on the face feature information corresponding to the center person image; and determining the set similarity between the plurality of second person image sets based on the center feature similarity.

[0122] Those skilled in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be realized by electronic hardware, computer software, or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been described in general terms in the above description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0123] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in random access memory (RAM), flash memory, read-only memory (ROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and

[0124] The above detailed description describes the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A character image clustering method, characterized in that: include: Obtain facial feature information corresponding to each person image in the person image collection; Determining the facial feature similarity between each person image and other person images based on the facial feature information; Clustering the person images whose facial feature similarity is greater than a first threshold to obtain a plurality of first person image sets; Optimizing and classifying the person images in each first person image set to obtain a plurality of second person image sets; comparing the set similarities between the plurality of second person image sets, clustering the second person image sets whose set similarities are greater than a second threshold, to obtain a third person image set; The optimizing and classifying the person images in each first person image set to obtain a plurality of second person image sets includes: determining a plurality of feature categories based on facial feature information corresponding to each person image in the first person image set; The person images in the first person image set are optimized and classified based on the feature categories to obtain multiple second person image sets.

2. The method according to claim 1, characterized in that The step of obtaining facial feature information corresponding to each person image in the person image set includes: Identifying a facial image corresponding to each person image in the person image set; Extract facial feature information from the facial image to obtain facial feature information corresponding to each person image in the person image set.

3. The method according to claim 2, characterized in that Clustering the person images whose facial feature similarity is greater than a first threshold to obtain a plurality of first person image sets includes: Based on the facial feature similarity between each character image and other character images, a similarity network structure diagram is constructed, wherein each node in the similarity network structure diagram is a face image, and the connecting line between each two nodes is the facial feature similarity; retaining the connection lines between every two nodes in the similarity network structure diagram whose corresponding facial feature similarity is greater than a first threshold, and deleting the remaining connection lines, to obtain multiple groups of connected person images; Each group of connected person images is clustered to obtain multiple first person image sets.

4. The method according to claim 3, characterized in that The set similarity between the plurality of second person image sets is obtained by: Obtaining facial feature similarities corresponding to any two person images in any two of the second person image set; The set similarity between the plurality of second person image sets is determined based on the facial feature similarity, where the set similarity is any one of average feature similarity, maximum feature similarity, and minimum feature similarity.

5. The method according to claim 3, characterized in that The set similarity between the plurality of second person image sets is obtained by: Obtaining a central person image corresponding to each second person image set; Determining the central feature similarity between the plurality of second person image sets based on the facial feature information corresponding to the central person image; The central feature similarity is used to determine the set similarity between the plurality of second person image sets.

6. A person image clustering device, characterized in that: include: An acquisition module is used to obtain facial feature information corresponding to each person image in the person image collection; A determination module, configured to determine the similarity of facial features between each person image and other person images based on the facial feature information; a clustering module, configured to cluster the person images whose facial feature similarity is greater than a first threshold value to obtain a plurality of first person image sets; The clustering module is further configured to optimize and classify the person images in each first person image set to obtain a plurality of second person image sets; The clustering module is further configured to compare the set similarities between the plurality of second person image sets, cluster the second person image sets whose set similarities are greater than a second threshold, and obtain a third person image set; The optimizing and classifying the person images in each first person image set to obtain a plurality of second person image sets includes: determining a plurality of feature categories based on facial feature information corresponding to each person image in the first person image set; The person images in the first person image set are optimized and classified based on the feature categories to obtain multiple second person image sets.

7. The person image clustering device according to claim 6, characterized in that: The acquisition module is specifically used to identify the face picture corresponding to each character image in the character image set; extract facial feature information from the face picture to obtain the facial feature information corresponding to each character image in the character image set.

8. A computer device, characterized in that: include: A processor and a memory, wherein the processor is configured to execute a character image clustering program stored in the memory to implement the character image clustering method according to any one of claims 1 to 5.

9. A storage medium, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the person image clustering method according to any one of claims 1 to 5.

Images