CN112818867B - Portrait clustering method, equipment and storage medium - Google Patents
Portrait clustering method, equipment and storage medium Download PDFInfo
- Publication number
- CN112818867B CN112818867B CN202110146066.1A CN202110146066A CN112818867B CN 112818867 B CN112818867 B CN 112818867B CN 202110146066 A CN202110146066 A CN 202110146066A CN 112818867 B CN112818867 B CN 112818867B
- Authority
- CN
- China
- Prior art keywords
- portrait
- suspected
- similarity
- same type
- face
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a portrait clustering method, equipment and a storage medium, wherein the portrait clustering method comprises the following steps: acquiring an image sequence; extracting a portrait picture in an image and portrait features in the portrait picture, and carrying out clustering treatment to obtain clustering results of different portrait features respectively; according to the relevance of different portrait features, fusing clustering results of the different portrait features to obtain the same type of portrait, suspected same type of portrait and different types of portrait; re-weighting the similarity of different portrait features of the suspected targets in the suspected same portrait to obtain the comprehensive similarity of the suspected targets; and comparing the comprehensive similarity with a preset threshold value, and classifying the suspected same type of portrait into the same type of portrait or different types of portrait. According to the application, the suspected objects of the suspected same-class images are subjected to secondary verification, so that the suspected same-class images are further and accurately classified, the data of the unable intermediate section are eliminated, and the recall rate of clustering is effectively increased.
Description
Technical Field
The application belongs to the technical field of video analysis, and particularly relates to a portrait clustering method, equipment and a storage medium.
Background
With the development of related technologies, face clustering is widely used. For example, a mobile phone album can gather face images of the same person into a group by a face clustering method; for example, when the public security industry breaks a case, the suspected person images without identity confirmation are required to be searched in a mass portrait library, and based on the searching, a one-person-one-file information library is required to be established, and the images of the same person in the information library belong to the same category.
In the prior art, images can be clustered by adopting a face clustering mode, however, the existing face clustering mode has low accuracy, so that more image data which cannot be classified is obtained, and the recall rate of a clustering method is low.
Disclosure of Invention
The application provides a portrait clustering method, equipment and a storage medium, which are used for solving the technical problem of low recall rate of the clustering method.
In order to solve the technical problems, the application adopts a technical scheme that: a portrait clustering method, comprising: acquiring an image sequence, wherein the image sequence comprises a plurality of frames of images; extracting a portrait picture in the image; extracting the portrait features in the portrait pictures, and respectively clustering different portrait features to obtain clustering results of the portrait features, wherein the clustering results comprise the same type of features, suspected same type of features and different types of features; according to the relevance of different portrait features, fusing the clustering results of the different portrait features to obtain the same type of portrait, suspected same type of portrait and different types of portrait; re-weighting the similarity of the characteristics of different figures of the suspected targets in the suspected same figure to obtain the comprehensive similarity of the suspected targets; and comparing the comprehensive similarity with a preset threshold value, and classifying the suspected same type of portrait into the same type of portrait or different types of portrait.
According to an embodiment of the present application, the portrait pictures include a face picture including a face code, a human body picture including a human face human body associated code, and a gait sequence picture including a human body gait associated code.
According to an embodiment of the present application, the extracting the portrait features in the portrait pictures, and clustering the portrait features respectively to obtain clustering results of the portrait features respectively, includes: extracting face features in the face picture, and carrying out clustering treatment on the face features to obtain a face clustering result, wherein the face clustering result comprises faces of the same type, suspected faces of the same type and faces of different types; extracting human body characteristics in the human body picture, and carrying out clustering treatment on the human body characteristics to obtain a human body clustering result, wherein the human body clustering result comprises the same type of human body, suspected same type of human body and different types of human body.
According to an embodiment of the present application, the fusing the clustering results of different portrait features according to the relevance of the portrait features to obtain the same type of portrait, suspected same type of portrait and different types of portrait includes: and according to the human face human body association code and the human gait association code, fusing the human face clustering result and the human body clustering result to obtain the same type of human images, suspected same type of human images and different types of human images.
According to an embodiment of the present application, the re-weighting the similarity of the features of the images of the suspected objects in the same suspected image includes: multiplying the face initial similarity of the suspected target in the suspected same type of portrait by a face weight value to obtain face similarity; multiplying the human body initial similarity of the suspected target by a human body weight value to obtain human body similarity; multiplying the gait initial similarity of the suspected target by a gait weight value to obtain gait similarity; adding the face similarity, the human body similarity and the gait similarity to obtain the comprehensive similarity; wherein the sum of the face weight value, the body weight value and the gait weight value is one.
According to an embodiment of the present application, the face weight value is greater than the gait weight value, and the gait weight value is greater than the body weight value.
According to an embodiment of the present application, the classifying the suspected identical person image into the identical person image or the different person images by comparing the integrated similarity with a preset threshold includes: judging whether the comprehensive similarity is greater than or equal to the preset threshold, wherein the preset threshold is the minimum similarity of the same type of figures; if the comprehensive similarity is greater than or equal to the preset threshold, the suspected same type of portrait belongs to the same type of portrait; and if the comprehensive similarity is smaller than the preset threshold, the suspected same-class portraits belong to different-class portraits.
According to an embodiment of the application, the method comprises: and storing the same type of figures in an archive database, and storing the different types of figures in a figure database.
In order to solve the technical problems, the application adopts another technical scheme that: an electronic device comprising a memory and a processor coupled to each other, the processor being configured to execute program instructions stored in the memory to implement any of the methods described above.
In order to solve the technical problems, the application adopts another technical scheme that: a computer readable storage medium having stored thereon program data which when executed by a processor implements any of the methods described above.
The beneficial effects of the application are as follows: compared with the prior art, the method and the device have the advantages that the suspected targets of the suspected images of the same class are subjected to secondary verification, so that the suspected images of the same class are further and accurately classified into the images of the same class or the images of different classes, data of middle sections which cannot be accurately classified are eliminated, and the recall rate of clustering is effectively increased.
Drawings
For a clearer description of the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly introduced below, it being obvious that the drawings in the description below are only some embodiments of the present application, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art, wherein:
FIG. 1 is a flow chart of an embodiment of a portrait clustering method according to the present application;
FIG. 2 is a schematic diagram of a portrait clustering apparatus according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a frame of an embodiment of an electronic device of the present application;
FIG. 4 is a schematic diagram of a frame of one embodiment of a computer-readable storage medium of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
Referring to fig. 1, an embodiment of the present application provides a method for clustering images, including the following steps:
S101: an image sequence is acquired, the image sequence comprising a plurality of frames of images.
And acquiring an image sequence, wherein the image sequence is a multi-frame image obtained by shooting by the shooting equipment. The image sequence may comprise consecutive multi-frame images, or the image sequence may also be discontinuous multi-frame images. Based on continuous multi-frame images, the method is more beneficial to acquiring and forming the moving track of the target.
The image capturing apparatus includes a general image capturing apparatus or an intelligent image capturing apparatus, which can perform image processing, and the general image capturing apparatus also needs to perform image processing in combination with an image processing module, such as a video stream processing module.
S102: and extracting the portrait picture in the image.
And extracting the portrait picture in the image. The portrait pictures comprise face pictures, human body pictures and gait sequence pictures. It should be noted that each image may include a face image or a human body image, or may include both a face image and a human body image, or may not include both a face image and a human body image. The gait sequence image is a portrait gait sequence obtained from a plurality of images, and the plurality of images are preferably continuous images.
The face picture comprises face codes, namely face tracking Id; the human body picture comprises a human face human body association code, namely a human face human body association Id; the gait sequence pictures comprise human gait related codes, namely human gait feature related ids. If the human body picture and the human face picture belong to the same person, the human body picture generates a human face human body association code which is the same as or can be associated with the human face picture according to the human face code; if the human body picture does not find the human face picture belonging to the same person, for example, the human body picture does not appear in the same image, the human body picture can have an independent code of the human body picture so as to obtain the human face human body association code after being further matched with the human face picture belonging to the same person. Similarly, if the gait sequence picture finds a human body picture which can be correlated with the human body picture, the gait sequence picture generates a human body gait correlation code which is the same as or can be correlated with the human body picture according to the code of the human body picture; if the gait sequence picture does not find the associable human body picture, the gait sequence picture can have an independent code of the gait sequence picture so as to obtain the human body gait association code after being further matched with the human body picture belonging to the same person.
It should be noted that, if the image capturing apparatus is an intelligent image capturing apparatus, a portrait picture in an image may be directly extracted. If the image capturing apparatus is a common image capturing apparatus, the image sequence is further input to an image processing module for image processing, for example, a video stream processing module, so as to extract a portrait picture in the image.
S103: and extracting the portrait features in the portrait pictures, and respectively clustering different portrait features to obtain clustering results of the different portrait features.
And extracting the portrait features in the portrait pictures, and respectively carrying out clustering treatment on different portrait features to obtain clustering results of the different portrait features, wherein the clustering results comprise the same type of features, suspected same type of features and different types of features.
The method specifically comprises the following steps:
Extracting face features in the face picture, and carrying out clustering processing on the face features to obtain a face clustering result, wherein the face clustering result comprises faces of the same type, suspected faces of the same type and faces of different types. The faces of the same class are a set of face pictures belonging to the same person; the suspected face of the same type is a set of face pictures of which the similarity reaches a preset range and suspected face pictures of the same person; and determining a set of face pictures which do not belong to the same person for different types of faces with similarity lower than a preset range.
The suspected target suspected to the face of the same type may be a target determined to be in the face of the same type, i.e. suspected to the face of the same type and the face of the same type. A suspected object suspected of being in the same class of face may not belong to an object determined to be in the same class of face.
Human body characteristics in the human body pictures are extracted, clustering processing is carried out on the human body characteristics, and human body clustering results are obtained, wherein the human body clustering results comprise the same type of human body, suspected same type of human body and different types of human body. The same type of human body is a collection of human body pictures belonging to the same person; human bodies of the same type are suspected to be a set of human body pictures of which the similarity reaches a preset range and suspected to belong to the same person; different human bodies are set of human body pictures with similarity lower than a preset range and not belonging to the same person.
The suspected target suspected of the same human body may be a target determined to be in the same human body, i.e. suspected of the same human body and the same human body. A suspected target suspected of being in the same class of human body may not belong to a target determined to be in the same class of human body.
S104: and fusing clustering results of different portrait features according to the relevance of the different portrait features to obtain the same type of portrait, suspected same type of portrait and different types of portrait.
The association of the portrait features includes the association of human face and human gait, and is represented by the association code of human face and human gait. And fusing the face clustering result and the human body clustering result through the face human body association codes, and fusing the same human face with the face human body association codes and the same human body to obtain the same human image. Different types of human faces and different types of human bodies are fused to obtain different types of human images.
It should be noted that the suspected face of the same type and the suspected human body of the same type can be temporarily classified into the suspected human body of the same type; or the suspected identical person images are considered as a virtual set, and the suspected identical person faces and the suspected identical person bodies are not actually classified into the suspected identical person images, but are directly classified into the identical person images or different person images after subsequent operation. The suspected faces and the suspected human bodies of the same class can be finally classified into the same class of human images or different class of human images.
In addition, if the same type of face is not matched with the corresponding same type of human body, the same type of human images are directly classified into the same type of human images. Similarly, if the same human body is not matched with the corresponding same human face, the same human body is also directly classified into the same human figure.
S105: and re-weighting the similarity of different portrait features of the suspected targets in the suspected same portrait to obtain the comprehensive similarity of the suspected targets.
Re-weighting the similarity of different portrait features of the suspected targets in the suspected same portrait, wherein the obtaining the comprehensive similarity of the suspected targets comprises the following steps:
Multiplying the face initial similarity (Fsimi ori) of the suspected target in the suspected same type of human image by a face weight value (P 1) to obtain the face similarity;
Multiplying the human body initial similarity (Psimi ori) of the suspected target by a human body weight value (P 2) to obtain human body similarity;
Multiplying the gait initial similarity (Gsimi ori) of the suspected target by a gait weight value (P 3) to obtain gait similarity; wherein the sum of the face weight value (P 1), the human body weight value (P 2) and the gait weight value (P 3) is one;
And adding the face similarity, the human body similarity and the gait similarity to obtain the comprehensive similarity (Simi new).
The specific calculation formula is as follows:
Siminew＝Fsimiori*P1+Psimiori*P2+Gsimiori*P3
If the suspected targets in the same type of portrait have no corresponding gait characteristics, the gait similarity is set to zero. The initial similarity of the image features is larger than or equal to a first preset value and smaller than a second preset value. For example simi ori e (85,100), the first predetermined value is 85, the second predetermined value is 100, although the first predetermined value may be 75, 80, 90, etc., without limitation.
In addition, P 1+P2+P3 =1, in general, the face features have the greatest effect on determining similarity, and gait features are inferior; thus, the human face weight value is larger than the gait weight value, and the gait weight value is larger than the human body weight value, for example, the human face weight value accounts for more than 60%, the gait weight value accounts for 20% -30%, and the human body weight value accounts for more than 10%. Of course, the face weight value, the human body weight value and the gait weight value can also be used for adjusting specific weights according to the corresponding face similarity, human body similarity and gait similarity.
S106: and comparing the comprehensive similarity with a preset threshold value, and classifying the suspected same type of portrait into the same type of portrait or different types of portrait.
The step of classifying the suspected same kind of portraits into the same kind of portraits or different kinds of portraits by comparing the comprehensive similarity with a preset threshold value comprises the following steps:
Judging whether the comprehensive similarity is greater than or equal to a preset threshold, wherein the preset threshold is the minimum similarity of the images belonging to the same class, and the preset threshold is an empirical value and can be adjusted according to experiments without limitation;
if the comprehensive similarity is greater than or equal to a preset threshold, the same portrait is suspected to belong to the same portrait;
if the comprehensive similarity is smaller than a preset threshold, the suspected same type of portrait belongs to different types of portraits.
If the suspected target suspected to the same type of portrait is a target determined to be in the same type of portrait, classifying the suspected same type of portrait into the corresponding same type of portrait when the comprehensive similarity is greater than or equal to a preset threshold. If the suspected targets suspected to the same type of portrait also do not belong to the targets determined to be in the same type of portrait, classifying the suspected same type of portrait as independent same type of portrait when the comprehensive similarity is greater than or equal to a preset threshold.
And carrying out secondary verification on the suspected targets suspected to be similar to the human images, so that the suspected to be similar to the human images are further and accurately classified to be similar to the human images or different human images, data of intermediate sections incapable of being accurately classified are eliminated, and the recall rate of clustering is effectively increased.
S107: the same kind of figures are stored in an archive database, and different kinds of figures are stored in a figure database.
After accurately classifying suspected identical human images into identical human images or different human images, storing the identical human images in an archive database for subsequent retrieval and the like; different types of portraits are stored in a portraits database for later rollback processing, namely, after more image resources exist, the different types of portraits can be further and accurately classified.
Referring to fig. 2, fig. 2 is a schematic diagram of a portrait clustering apparatus according to an embodiment of the present application.
Still another embodiment of the present application provides a portrait clustering apparatus, which includes an acquisition module 21 and a processing module 22. The acquisition module 21 acquires an image sequence including a plurality of frames of images. The processing module 22 extracts a portrait picture in the image; the processing module 22 extracts the portrait features in the portrait pictures, and respectively performs clustering processing on different portrait features to respectively obtain clustering results of the different portrait features, wherein the clustering results comprise the same type of features, suspected same type of features and different types of features; the processing module 22 fuses the clustering results of the different portrait features according to the relevance of the different portrait features to obtain the same type of portrait, suspected same type of portrait and different types of portrait; the processing module 22 re-weights the similarity of the characteristics of different figures of the suspected targets in the suspected same figure to obtain the comprehensive similarity of the suspected targets; the processing module 22 classifies the suspected same type of portrait with the same type of portrait or different types of portrait by comparing the integrated similarity with a preset threshold.
The device carries out secondary verification on the suspected targets suspected to the same type of human images, so that the suspected same type of human images are further and accurately classified to the same type of human images or different types of human images, data of middle sections incapable of being accurately classified are eliminated, and the recall rate of clustering is effectively increased.
Referring to fig. 3, fig. 3 is a schematic frame diagram of an embodiment of an electronic device according to the present application.
Still another embodiment of the present application provides an electronic device 30, including a memory 31 and a processor 32 coupled to each other, where the processor 32 is configured to execute program instructions stored in the memory 31 to implement the image clustering method of any one of the above embodiments. In one particular implementation scenario, electronic device 30 may include, but is not limited to: the microcomputer and the server, and the electronic device 30 may also include a mobile device such as a notebook computer and a tablet computer, which is not limited herein.
Specifically, the processor 32 is configured to control itself and the memory 31 to implement the steps in the portrait clustering method of any of the embodiments described above. The processor 32 may also be referred to as a CPU (Central Processing Unit ). The processor 32 may be an integrated circuit chip having signal processing capabilities. The Processor 32 may also be a general purpose Processor, a digital signal Processor (DIGITAL SIGNAL Processor, DSP), an Application SPECIFIC INTEGRATED Circuit (ASIC), a Field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, a discrete gate or transistor logic device, a discrete hardware component. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. In addition, the processor 32 may be commonly implemented by an integrated circuit chip.
Referring to fig. 4, fig. 4 is a schematic diagram of a frame of an embodiment of a computer readable storage medium according to the present application.
A further embodiment of the present application provides a computer readable storage medium 40 having stored thereon program data 41, which when executed by a processor, implements the portrait clustering method of any of the embodiments described above.
In the several embodiments provided in the present application, it should be understood that the disclosed method and apparatus may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of modules or units is merely a logical functional division, and there may be additional divisions of actual implementation, e.g., units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical, or other forms.
The elements illustrated as separate elements may or may not be physically separate, and elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over network elements. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium 40. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a readable storage medium 40, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the method of the embodiments of the present application. And the aforementioned readable storage medium 40 includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The foregoing description is only illustrative of the present application and is not intended to limit the scope of the application, and all equivalent structures or equivalent processes or direct or indirect application in other related technical fields are included in the scope of the present application.
Claims (10)
1. A method of human image clustering, comprising:
Acquiring an image sequence, wherein the image sequence comprises a plurality of frames of images;
extracting a portrait picture in the image, wherein the portrait picture comprises a face picture, a human body picture and a gait sequence picture;
extracting the portrait features in the portrait pictures, and respectively clustering different portrait features to obtain clustering results of the portrait features, wherein the clustering results comprise the same type of features, suspected same type of features and different types of features;
According to the relevance of different portrait features, fusing the clustering results of the different portrait features to obtain the same type of portrait, suspected same type of portrait and different types of portrait;
Re-weighting the similarity of different portrait features of the suspected targets in the suspected same portrait to obtain the comprehensive similarity of the suspected targets, wherein the similarity of the different portrait features comprises initial human face similarity, initial human body similarity and initial gait similarity;
and comparing the comprehensive similarity with a preset threshold value, and classifying the suspected same type of portrait into the same type of portrait or different types of portrait.
2. The method of claim 1, wherein the face picture comprises a face code, the body picture comprises a face-body associated code, and the gait sequence picture comprises a body gait associated code.
3. The method according to claim 2, wherein the extracting the portrait features in the portrait pictures, and clustering the portrait features respectively, so as to obtain clustering results of the portrait features respectively, includes:
Extracting face features in the face picture, and carrying out clustering treatment on the face features to obtain a face clustering result, wherein the face clustering result comprises faces of the same type, suspected faces of the same type and faces of different types;
Extracting human body characteristics in the human body picture, and carrying out clustering treatment on the human body characteristics to obtain a human body clustering result, wherein the human body clustering result comprises the same type of human body, suspected same type of human body and different types of human body.
4. The method of claim 3, wherein fusing the clustering results of the different portrait features according to the relevance of the different portrait features to obtain the same type of portrait, suspected same type of portrait, and different types of portrait comprises:
and according to the human face human body association code and the human gait association code, fusing the human face clustering result and the human body clustering result to obtain the same type of human images, suspected same type of human images and different types of human images.
5. The method of claim 4, wherein the re-weighting the similarity of the features of the different images of the suspected objects in the same class of images to obtain the integrated similarity of the suspected objects comprises:
Multiplying the face initial similarity of the suspected target in the suspected same type of portrait by a face weight value to obtain face similarity;
Multiplying the human body initial similarity of the suspected target by a human body weight value to obtain human body similarity;
Multiplying the gait initial similarity of the suspected target by a gait weight value to obtain gait similarity;
adding the face similarity, the human body similarity and the gait similarity to obtain the comprehensive similarity;
Wherein the sum of the face weight value, the body weight value and the gait weight value is one.
6. The method of claim 5, wherein the face weight value is greater than the gait weight value, the gait weight value being greater than the body weight value.
7. The method of claim 1, wherein classifying the suspected identical class of portraits into the identical class of portraits or the different class of portraits using the integrated similarity to a preset threshold comprises:
Judging whether the comprehensive similarity is greater than or equal to the preset threshold, wherein the preset threshold is the minimum similarity of the same type of figures;
if the comprehensive similarity is greater than or equal to the preset threshold, the suspected same type of portrait belongs to the same type of portrait;
and if the comprehensive similarity is smaller than the preset threshold, the suspected same-class portraits belong to different-class portraits.
8. The method according to claim 1, characterized in that the method comprises:
and storing the same type of figures in an archive database, and storing the different types of figures in a figure database.
9. An electronic device comprising a memory and a processor coupled to each other, the processor being configured to execute program instructions stored in the memory to implement the method of any one of claims 1 to 8.
10. A computer readable storage medium having stored thereon program data, which when executed by a processor implements the method of any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110146066.1A CN112818867B (en) | 2021-02-02 | 2021-02-02 | Portrait clustering method, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110146066.1A CN112818867B (en) | 2021-02-02 | 2021-02-02 | Portrait clustering method, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112818867A CN112818867A (en) | 2021-05-18 |
| CN112818867B true CN112818867B (en) | 2024-05-31 |
Family
ID=75860674
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110146066.1A Active CN112818867B (en) | 2021-02-02 | 2021-02-02 | Portrait clustering method, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112818867B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113360688B (en) * | 2021-06-28 | 2024-02-20 | 北京百度网讯科技有限公司 | Method, device and system for constructing information base |
| CN113657430B (en) * | 2021-06-30 | 2025-02-21 | 浙江大华技术股份有限公司 | Pedestrian clustering method, device and storage medium |
| CN113657434B (en) * | 2021-07-02 | 2024-12-06 | 浙江大华技术股份有限公司 | Face and body association method, system and computer-readable storage medium |
| CN113469095B (en) * | 2021-07-13 | 2023-05-16 | 浙江大华技术股份有限公司 | Gait-based character secondary verification method and device |
| CN114663954B (en) * | 2022-03-30 | 2024-08-13 | 济南博观智能科技有限公司 | Clustering method, system and device for face images and computer readable storage medium |
| CN115953650B (en) * | 2023-03-01 | 2023-06-27 | 杭州海康威视数字技术股份有限公司 | Training method and device for feature fusion model |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103902961A (en) * | 2012-12-28 | 2014-07-02 | 汉王科技股份有限公司 | Face recognition method and device |
| WO2015089949A1 (en) * | 2013-12-19 | 2015-06-25 | 成都品果科技有限公司 | Human face clustering method merging lbp and gabor features |
| WO2017088470A1 (en) * | 2015-11-27 | 2017-06-01 | 小米科技有限责任公司 | Image classification method and device |
| CN109783685A (en) * | 2018-12-28 | 2019-05-21 | 上海依图网络科技有限公司 | A kind of querying method and device |
| JP2019109760A (en) * | 2017-12-19 | 2019-07-04 | セコム株式会社 | Image monitoring device |
| CN110413815A (en) * | 2019-07-26 | 2019-11-05 | 北京旷视科技有限公司 | Portrait clusters cleaning method and device |
| CN110569777A (en) * | 2019-08-30 | 2019-12-13 | 深圳市商汤科技有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN110765863A (en) * | 2019-09-17 | 2020-02-07 | 清华大学 | Target clustering method and system based on space-time constraint |
| CN110874310A (en) * | 2018-12-21 | 2020-03-10 | 北京安天网络安全技术有限公司 | Terminal behavior monitoring method and device, electronic equipment and storage medium |
| CN112131419A (en) * | 2020-08-17 | 2020-12-25 | 浙江大华技术股份有限公司 | Image archive merging method and device, electronic equipment and storage medium |
| CN112270310A (en) * | 2020-11-24 | 2021-01-26 | 上海工程技术大学 | Cross-camera pedestrian multi-target tracking method and device based on deep learning |
| CN112269889A (en) * | 2020-09-23 | 2021-01-26 | 上海市刑事科学技术研究院 | Interactive method, client and system for searching difficult portrait |
-
2021
- 2021-02-02 CN CN202110146066.1A patent/CN112818867B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103902961A (en) * | 2012-12-28 | 2014-07-02 | 汉王科技股份有限公司 | Face recognition method and device |
| WO2015089949A1 (en) * | 2013-12-19 | 2015-06-25 | 成都品果科技有限公司 | Human face clustering method merging lbp and gabor features |
| WO2017088470A1 (en) * | 2015-11-27 | 2017-06-01 | 小米科技有限责任公司 | Image classification method and device |
| JP2019109760A (en) * | 2017-12-19 | 2019-07-04 | セコム株式会社 | Image monitoring device |
| CN110874310A (en) * | 2018-12-21 | 2020-03-10 | 北京安天网络安全技术有限公司 | Terminal behavior monitoring method and device, electronic equipment and storage medium |
| CN109783685A (en) * | 2018-12-28 | 2019-05-21 | 上海依图网络科技有限公司 | A kind of querying method and device |
| CN110413815A (en) * | 2019-07-26 | 2019-11-05 | 北京旷视科技有限公司 | Portrait clusters cleaning method and device |
| CN110569777A (en) * | 2019-08-30 | 2019-12-13 | 深圳市商汤科技有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN110765863A (en) * | 2019-09-17 | 2020-02-07 | 清华大学 | Target clustering method and system based on space-time constraint |
| CN112131419A (en) * | 2020-08-17 | 2020-12-25 | 浙江大华技术股份有限公司 | Image archive merging method and device, electronic equipment and storage medium |
| CN112269889A (en) * | 2020-09-23 | 2021-01-26 | 上海市刑事科学技术研究院 | Interactive method, client and system for searching difficult portrait |
| CN112270310A (en) * | 2020-11-24 | 2021-01-26 | 上海工程技术大学 | Cross-camera pedestrian multi-target tracking method and device based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112818867A (en) | 2021-05-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112818867B (en) | Portrait clustering method, equipment and storage medium | |
| CN110489951B (en) | Risk identification method and device, computer equipment and storage medium | |
| CN112232148B (en) | Image clustering, target track tracking method, electronic device and storage medium | |
| CN110781733B (en) | Image duplicate removal method, storage medium, network equipment and intelligent monitoring system | |
| EP3779775B1 (en) | Media processing method and related apparatus | |
| CN112016623B (en) | Face clustering method, device, equipment and storage medium | |
| JP2022518469A (en) | Information processing methods and devices, storage media | |
| CN111104925B (en) | Image processing method, image processing apparatus, storage medium, and electronic device | |
| CN110941978B (en) | Face clustering method and device for unidentified personnel and storage medium | |
| Li et al. | Cross-level parallel network for crowd counting | |
| CN112052251B (en) | Target data updating method and related device, equipment and storage medium | |
| CN111985360A (en) | Face recognition method, device, equipment and medium | |
| CN113657430B (en) | Pedestrian clustering method, device and storage medium | |
| CN112307934A (en) | Image detection method, and training method, device, equipment and medium of related model | |
| CN113901981A (en) | Equipment clustering method and device, computer equipment and storage medium | |
| CN109800318A (en) | A kind of archiving method and device | |
| CN111783743A (en) | Image clustering method and device | |
| CN114841705A (en) | Anti-fraud monitoring method based on scene recognition | |
| CN112101483B (en) | Image clustering method, device electronic equipment and computer readable storage medium | |
| CN115565222A (en) | Face recognition method, face recognition system, terminal device and storage medium | |
| CN113918510A (en) | Picture archiving method and device, terminal equipment and computer readable storage medium | |
| CN114926973B (en) | Video monitoring method, device, system, server and readable storage medium | |
| CN113887366A (en) | Identity archiving method and device, electronic equipment and storage medium | |
| CN117369637A (en) | Product pushing method, device, equipment and medium for virtual reality | |
| CN115766215A (en) | Abnormal flow detection method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |