CN110147455B - Face matching retrieval device and method - Google Patents
Face matching retrieval device and method Download PDFInfo
- Publication number
- CN110147455B CN110147455B CN201710760859.6A CN201710760859A CN110147455B CN 110147455 B CN110147455 B CN 110147455B CN 201710760859 A CN201710760859 A CN 201710760859A CN 110147455 B CN110147455 B CN 110147455B
- Authority
- CN
- China
- Prior art keywords
- face
- index
- features
- face features
- library
- 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
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000004422 calculation algorithm Methods 0.000 claims description 27
- 230000008569 process Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 210000000887 face Anatomy 0.000 description 18
- 238000004364 calculation method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000001815 facial effect Effects 0.000 description 3
- 238000005192 partition Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Library & Information Science (AREA)
- Physics & Mathematics (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Processing Or Creating Images (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a face matching retrieval device and a face matching retrieval method, and relates to the fields of image processing, matching, retrieval and the like. The invention discloses a face matching retrieval device, which at least comprises an index comparison server and a secondary comparison server, wherein: the index comparison server is used for comparing the face features of the face images to be retrieved with the face features in a pre-established index library to obtain corresponding indexes, wherein each index in the index library corresponds to one type of face features respectively; and the secondary comparison server is used for searching the face features of the face image to be searched in all the face features stored in the corresponding area in the pre-established face feature database in the index obtained by the index comparison server, and obtaining a comparison result.
Description
Technical Field
The invention relates to the fields of image processing, matching, searching and the like, in particular to a scheme for carrying out quick matching searching in a large number of face images.
Background
Along with the development of image processing and pattern recognition technology, the application of face recognition and comparison is wider and wider. Particularly, along with the development of video big data, a face image is often required to be input in the public security field, and then a face matched with the face image is found out from a massive face library. Face libraries in the public security field often contain millions or even tens of millions of face images. How to quickly and accurately match and retrieve faces from a large number of face libraries becomes a current urgent need to be solved.
At present, a method for quickly matching massive human faces generally adopts a distributed parallel processing mode, massive human face comparison tasks are integrated into zero by adding machines, the zero is evenly distributed to each machine for processing, and then the comparison results are combined.
Disclosure of Invention
The device and the method for matching and searching the human face can solve the problem of low efficiency of the existing human face searching.
The invention discloses a face matching retrieval device, which at least comprises an index comparison server and a secondary comparison server, wherein:
the index comparison server is used for comparing the face features of the face images to be retrieved with the face features in a pre-established index library to obtain corresponding indexes, wherein each index in the index library corresponds to one type of face features respectively;
and the secondary comparison server is used for searching the face features of the face image to be searched in all the face features stored in the corresponding area in the pre-established face feature database in the index obtained by the index comparison server, and obtaining a comparison result.
Optionally, the apparatus further includes:
the feature management server is used for performing management operation on the face feature database;
the management operations include any one or more of the following: adding face images, deleting face images, modifying face images, and querying face images.
Optionally, in the above device, the index comparison server is configured to compare, when a new face image to be put in storage is to be obtained, the new face image to be put in storage with face features in the index library, so as to obtain a corresponding index;
and the feature management server is used for storing the newly added face image to be put in storage into the region corresponding to the index in the face feature database according to the obtained index.
The invention also discloses a face matching retrieval method, which comprises the following steps:
comparing the face features of the face images to be retrieved with the face features in a pre-established index library to obtain corresponding indexes;
searching the face features of the face image to be searched in the face features stored in the region corresponding to the obtained index in a pre-established face feature database to obtain a comparison result;
wherein, each index in the index library corresponds to a type of face feature respectively.
Optionally, before comparing the face features of the face image to be retrieved with the face features in the index library established in advance, the method further includes:
and finding out various face features from the face images obtained in advance according to a clustering method to serve as indexes, and establishing an index library.
Optionally, in the above method, comparing the face features of the face image to be retrieved with the face features in the pre-established index library to obtain the corresponding index includes:
comparing the face features of the face images to be retrieved with the face features in a pre-established index library by adopting a distance calculation algorithm, and taking the index with the similarity in the comparison result in a set range as a corresponding index;
the algorithm for calculating the distance is the same as the algorithm for calculating the distance by clustering adopted for generating the index when the index library is built in advance.
Optionally, in the above method, in the pre-established face feature database, the searching the face features of the face image to be searched in the face features stored in the area corresponding to the obtained index includes:
searching the face features of the face image to be searched in the face features stored in the region corresponding to the obtained index by adopting a distance calculation algorithm;
the algorithm for calculating the distance is the same as the algorithm for calculating the distance by clustering adopted for generating the index when the index library is built in advance.
Optionally, the method further comprises:
performing management operations on the face feature database, wherein the management operations comprise any one or more of the following operations:
adding face images, deleting face images, modifying face images, and querying face images.
Optionally, in the above method, when the management operation performed on the face feature database is a new face image, the method includes:
comparing the newly added face image to be put in storage with face features in the index library to obtain a corresponding index;
and storing the newly added face image to be put in storage into a region corresponding to the index in the face feature database according to the obtained index.
Also disclosed herein is a retrieval device comprising a memory, a processor and a computer program stored on said memory and executable on said processor, characterized in that said processor implements the processing of the method as described above when executing said computer program.
Compared with the prior art, the technical scheme of the application has the following beneficial effects:
by establishing indexes, the mass face features are roughly classified (namely, one index corresponds to one type of face feature), so that the types of the face features are determined firstly during face retrieval, the range to be matched can be greatly reduced, and the recognition speed is improved;
in addition, by establishing the index, only the index is compared with the representative face features (namely indexes) with limited quantity, so that the number of times of matching is reduced, the computing resource is saved, and the hardware cost is reduced;
the recognition speed is improved, waiting time of a user in face retrieval in a mass face library is reduced, and user experience and working efficiency of the user are improved.
Drawings
FIG. 1 is a schematic diagram of a face quick matching search device in an embodiment of the invention;
FIG. 2 is a flowchart of a face quick matching retrieval method in an embodiment of the invention;
FIG. 3 is a schematic diagram of a rapid matching retrieval device for a large number of faces in an alternative embodiment of the present invention;
FIG. 4 is a process flow diagram of an index comparison server in an alternative embodiment of the invention;
FIG. 5 is a flow chart of generating an index in an alternative embodiment of the invention;
FIG. 6 is a flow chart of the face feature data binning in an alternative embodiment of the present invention;
fig. 7 is a complete flow chart of face feature retrieval in an alternative embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be described in further detail with reference to the specific embodiments. It should be noted that, in the case of no conflict, the embodiments of the present application and the features in the embodiments may be arbitrarily combined with each other.
The inventor of the application proposes that face indexes are built in a massive face library in a clustering mode, when face retrieval is carried out, faces to be retrieved are compared with indexes with limited quantity, then the face range needing secondary comparison is obtained by utilizing the indexes, and secondary matching is carried out in a small range, so that the face retrieval efficiency is improved.
Based on the above-mentioned thought, this embodiment provides a complete search device for fast matching in a large number of faces, as shown in fig. 1, mainly including: index comparison server, secondary comparison server,
And the index comparison server is used for obtaining an index corresponding to the face feature to be processed by matching the face feature to be processed with the face feature in the index library. Wherein each index in the index library corresponds to a type of face feature. The face features are representative features. Representative features referred to herein may be local features of a face, such as features of eyes, mouth, facial shape, etc., as well as combined features of a face, such as any combination between various local features of a face, etc. The mass human faces can be divided into multiple types of human faces according to different representative characteristics. In practical application, a clustering method can be adopted to cluster faces with different representative characteristics from a large number of faces. And the secondary comparison server is used for carrying out secondary comparison and retrieval of the face features in the face range corresponding to the index calculated by the index server in the face feature database (namely, the face features stored in the area corresponding to the index) in the face feature recognition process.
In practical application, the index comparison server may have the following two application scenarios in the present device:
1) When the face feature data is put into storage, firstly, the index with the highest similarity with the face feature to be put into storage is obtained through face feature comparison, and then the index is stored into a corresponding partition of the face feature database;
2) When searching the human face, comparing the human face characteristics to be searched with the human face characteristics in the index library to obtain one or a plurality of indexes meeting a certain similarity. When in processing, the index comparison server firstly loads indexes from an index database to a calculation buffer or special hardware for index comparison, such as FPGA, GPU and the like; and comparing the extracted features of the face image to be processed with features in an index library to obtain corresponding indexes.
It is worth noting that the index comparison server calculates the distance algorithm needed by the feature comparison of the faces to be searched in the index library, the clustering distance calculation mode adopted when the index library is established to generate the index, and the secondary comparison server keeps the same with the distance calculation algorithm when the secondary comparison server performs the secondary comparison of the faces to be searched.
The index library stores all indexes adopted in the system, and each index referred to in the document corresponds to a group of representative face features (namely face features of the same type) clustered in a massive face through a clustering method.
The face feature database can store all face feature data adopted in the system according to index partition, wherein the face feature database is divided into areas corresponding to the indexes respectively, and all face features of one type of face feature corresponding to the index are stored in the areas corresponding to the indexes. For example, the face features stored in the regions corresponding to the respective indexes include all face feature data having a similarity of representative face features corresponding to the indexes within a set similarity threshold.
In addition, the device can further comprise a feature management server for managing the face feature data. The method can comprise the operations of adding, deleting, modifying, searching, obtaining corresponding face feature data according to indexes and the like of a face feature database. The secondary comparison server performs secondary retrieval on the face image to be retrieved, which can be realized through the feature management server. For example, when the secondary comparison server performs secondary search, all face feature data corresponding to the index in the face feature database may be loaded into the local or cache space of the secondary comparison server in real time or in advance by the feature management server, so as to perform secondary search. The index comparison server, the secondary comparison server, the feature management server, the index library and the feature library. The five parts can be respectively deployed on different machines according to functions, and can also be deployed on the same machine or a plurality of machines according to required computing resources.
In addition to the above-mentioned face image searching operation, there is a new face image warehousing operation, that is, when the index comparison server is waiting for the face image to be warehoused, the face image to be warehoused is compared with the face features in the index library, so as to obtain the corresponding index. And then, the feature management server stores the face image to be put in storage into a partition corresponding to the index in the face feature database according to the obtained index.
The embodiment also provides a method for quickly matching and searching massive faces, which is shown in fig. 2 and comprises the following operations:
step 200: searching the face features of the face images to be searched and the face features in a pre-established index library to obtain corresponding indexes;
wherein each index in the index library corresponds to a group of representative face features (namely, a class of face features);
in the step, the index corresponding to the search result comprises one or a plurality of indexes meeting a certain similarity.
For example, an algorithm for calculating a distance may be adopted, face features of the face image to be retrieved are compared with face features in a pre-established index library, and one or more indexes with similarity in a set range in the comparison result are used as corresponding indexes. The algorithm for calculating the distance is the same as the algorithm for calculating the distance by clustering adopted for generating the index when the index library is built in advance.
It is noted that representative features referred to herein may be local features of a face, such as features of eyes, mouth, facial shapes, etc., or may be combined features of a face, such as any combination between various local features of a face, etc. The mass human faces can be divided into multiple types of human faces according to different representative characteristics. In practical application, a clustering method can be adopted to cluster faces with different representative characteristics from a large number of faces.
Step 201: and carrying out secondary retrieval on the face features of the face image to be retrieved in all face feature frames stored in the corresponding areas in the pre-established face feature database in the obtained index, and obtaining a comparison result.
The face features corresponding to the indexes in the face feature database comprise all face feature data, wherein the similarity of the representative face features corresponding to the indexes (namely, the face features corresponding to the indexes) is within a set similarity threshold.
In the face features stored in the region corresponding to the obtained index, the face features of the face image to be searched can be searched by adopting an algorithm for calculating the distance. The algorithm for calculating the distance is the same as the algorithm for calculating the distance by clustering adopted for generating the index when the index library is built in advance.
On the basis of the above operation, the face feature database may be further managed, and the management operation performed may include any one or more of the following operations:
adding face images, deleting face images, modifying face images, and querying face images.
Taking the operation of adding a new face image as an example, the process is as follows:
comparing the newly added face image to be put in storage with face features in the index library to obtain a corresponding index;
and storing the newly added face image to be put in storage into a region corresponding to the index in the face feature database according to the obtained index.
In addition, there is provided a retrieval device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the processing of the method as described above when executing the computer program.
The following describes the implementation of the above scheme with reference to the drawings.
The structure diagram of the rapid matching and searching device for massive faces is shown in fig. 3, and the device comprises: the system comprises an index comparison server, a secondary comparison server, a feature management server, an index database and a face feature database.
The index library stores index information adopted in the system, wherein the stored indexes are representative face features clustered in a mass of faces by a clustering method (a group of representative face features are a class of face features). The index comparison server is mainly used for calculating indexes corresponding to the given face features; when a face is put in storage, an index comparison server calculates an index which is the closest to the face feature to be put in storage, and stores the related information of the face feature to be put in storage into a region corresponding to the closest index in a face feature database; when searching the human face, the index comparison server calculates all indexes which accord with a certain similarity threshold value with the human face to be searched, and matching search of the human face is carried out by utilizing the indexes through the secondary comparison server. The secondary comparison server obtains the face features to be compared through the feature management server by utilizing the index calculated by the index comparison server, and compares the face features to be searched with the obtained face features to obtain a comparison result. The feature management server is used for managing the face features, and specifically comprises the following steps: and adding, deleting, modifying, checking and indexing the face feature database to obtain corresponding face data.
It should be noted that, for ease of understanding, in the present structure diagram, the number of index comparison servers, secondary comparison servers, feature management servers, index libraries, and feature libraries is only one, but the actual scheme is not limited to one, and an appropriate number may be configured according to the traffic, and of course, the three servers may be on the same machine, or may be three different functional modules, processes, or threads instead of a single service.
The process flow of the index comparison server at the time of retrieval, as shown in fig. 4, includes the steps of:
step S401: the index comparison server loads the index into memory or hardware (e.g., FPGA, GPU, etc.) for comparison.
Step S402: and comparing the face features of the extracted face images to be processed with the face features in the index library to obtain corresponding indexes.
In this step, when comparing with features in the index library, the distance calculation algorithm needs to be consistent with the distance calculation method in the clusters used in the generation of the index and the distance calculation algorithm in the feature comparison in the secondary comparison.
When the face features of the face image to be processed are compared with the face features in the index library, the similarity of the representative face features corresponding to the index is within a set similarity threshold, and the face features of the face image to be processed are considered to correspond to the index. It can also be seen that the index corresponding to the face image to be processed obtained by comparison may be one or a plurality of indexes.
An indexing flow is generated, as shown in fig. 5, which includes the steps of:
step S501: extracting massive face features;
in this step, in order to obtain optimal retrieval accuracy and performance, the face corresponding to the extracted face feature is preferably a face in a face library to be retrieved later.
Step S502: and finding out representative face features by a clustering method to serve as indexes.
It should be noted that the distance calculation method in the cluster is consistent with the distance calculation method in the feature comparison.
The face feature data warehouse-in process, as shown in fig. 6, includes the following steps:
step S601: and extracting the face characteristics of the face to be put in storage.
Step S602: and comparing the similarity between the face features to be put in the warehouse and the features in the index warehouse to obtain the most similar index.
Step S603-1: and (4) inserting the face features to be put in the database into the corresponding region of the most similar index in the face feature database calculated in step S402.
Step S603-2: and updating the face characteristic data loaded by the characteristic management server.
Based on the above operations, the complete process of face feature retrieval, as shown in fig. 7, includes the following steps:
step S701: in the data initialization stage, the index comparison server loads an index library into a memory or hardware for comparison, and the feature management server loads feature data in a face feature database into the memory or hardware for comparison;
hardware for comparison, such as: FPGA, GPU, etc.
Step S702: and in the index comparison stage, the face features of the face pictures to be retrieved are compared with all indexes on an index server to obtain one or a plurality of indexes meeting the conditions.
In this step, the comparison may be implemented in the CPU by a comparison algorithm implemented by software code, or may be implemented by hardware such as FPGA and GPU in order to obtain higher performance.
When comparing with all indexes, indexes with similarity of facial features of the face picture to be retrieved within a range of a preset similarity threshold value can be considered as indexes meeting the conditions.
Step S703: and in the secondary comparison stage, the secondary comparison server performs secondary retrieval on the characteristics of the face image to be retrieved in the face characteristic range corresponding to the index obtained in the step S702, and a comparison result is obtained.
In this step, the comparison may be implemented in the CPU by a comparison algorithm implemented by software code, or may be implemented by hardware such as FPGA and GPU in order to obtain higher performance.
Step S704: and in the result output stage, the page acquires and displays the search result.
The comparison result obtained in step S703 may be stored in a temporary table or a local file of a memory or a database, for displaying a page.
Those of ordinary skill in the art will appreciate that all or a portion of the steps of the methods described above may be implemented by a program that instructs associated hardware, and the program may be stored on a computer readable storage medium such as a read-only memory, a magnetic or optical disk, etc. Alternatively, all or part of the steps of the above embodiments may be implemented using one or more integrated circuits. Accordingly, each module/unit in the above embodiment may be implemented in the form of hardware, or may be implemented in the form of a software functional module. The present application is not limited to any specific form of combination of hardware and software.
The foregoing is merely a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. The face matching and searching device at least comprises an index comparison server and a secondary comparison server, wherein:
the index comparison server is used for comparing the face features of the face images to be retrieved with the face features in a pre-established index library to obtain corresponding indexes, wherein each index in the index library corresponds to one type of face features respectively;
the secondary comparison server is used for searching the face features of the face images to be searched in all face features stored in the corresponding areas in the pre-established face feature database in the index obtained by the index comparison server, and obtaining a comparison result;
the index comparison server is further used for finding out various face features from the pre-acquired face images to serve as indexes according to a clustering method before comparing the face features of the face images to be retrieved with the face features in the pre-established index library, and establishing the index library;
the step of comparing the face features of the face images to be retrieved with the face features in the pre-established index library to obtain corresponding indexes comprises the following steps:
comparing the face features of the face images to be retrieved with the face features in a pre-established index library by adopting a distance calculation algorithm, and taking the index with the similarity in the comparison result in a set range as a corresponding index;
the algorithm for calculating the distance is the same as the algorithm for calculating the distance by clustering adopted for generating the index when the index library is built in advance.
2. The apparatus as recited in claim 1, further comprising:
the feature management server is used for performing management operation on the face feature database;
the management operations include any one or more of the following: adding face images, deleting face images, modifying face images, and querying face images.
3. The apparatus of claim 2, wherein,
the index comparison server is used for comparing the newly added face image to be put into storage with the face features in the index library when the newly added face image to be put into storage is needed, so as to obtain a corresponding index;
and the feature management server is used for storing the newly added face image to be put in storage into the region corresponding to the index in the face feature database according to the obtained index.
4. A face matching retrieval method comprises the following steps:
comparing the face features of the face images to be retrieved with the face features in a pre-established index library to obtain corresponding indexes;
searching the face features of the face image to be searched in the face features stored in the region corresponding to the obtained index in a pre-established face feature database to obtain a comparison result;
wherein each index in the index library corresponds to a type of face features respectively;
before comparing the face features of the face images to be retrieved with the face features in the index library established in advance, the method further comprises the following steps:
according to a clustering method, various face features are found out from a pre-acquired face image to serve as indexes, and an index library is established;
the step of comparing the face features of the face images to be retrieved with the face features in the pre-established index library to obtain corresponding indexes comprises the following steps:
comparing the face features of the face images to be retrieved with the face features in a pre-established index library by adopting a distance calculation algorithm, and taking the index with the similarity in the comparison result in a set range as a corresponding index;
the algorithm for calculating the distance is the same as the algorithm for calculating the distance by clustering adopted for generating the index when the index library is built in advance.
5. The method according to claim 4, wherein the searching the face features of the face image to be searched among the face features stored in the area corresponding to the obtained index in the pre-established face feature database includes:
searching the face features of the face image to be searched in the face features stored in the region corresponding to the obtained index by adopting a distance calculation algorithm;
the algorithm for calculating the distance is the same as the algorithm for calculating the distance by clustering adopted for generating the index when the index library is built in advance.
6. The method as recited in claim 5, further comprising:
performing management operations on the face feature database, wherein the management operations comprise any one or more of the following operations:
adding face images, deleting face images, modifying face images, and querying face images.
7. The method of claim 6, wherein the managing the face feature database when the face feature database is a newly added face image comprises:
comparing the newly added face image to be put in storage with face features in the index library to obtain a corresponding index;
and storing the newly added face image to be put in storage into a region corresponding to the index in the face feature database according to the obtained index.
8. A retrieval device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the process of the method of any of claims 4-7 when the computer program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710760859.6A CN110147455B (en) | 2017-08-30 | 2017-08-30 | Face matching retrieval device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710760859.6A CN110147455B (en) | 2017-08-30 | 2017-08-30 | Face matching retrieval device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110147455A CN110147455A (en) | 2019-08-20 |
| CN110147455B true CN110147455B (en) | 2024-03-22 |
Family
ID=67588060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710760859.6A Active CN110147455B (en) | 2017-08-30 | 2017-08-30 | Face matching retrieval device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110147455B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112559789A (en) * | 2019-09-10 | 2021-03-26 | 中兴通讯股份有限公司 | Face searching method, device, equipment and computer readable storage medium |
| CN110765292A (en) * | 2019-10-24 | 2020-02-07 | 重庆紫光华山智安科技有限公司 | Image retrieval method, training method and related device |
| CN110874419B (en) * | 2019-11-19 | 2022-03-29 | 山东浪潮科学研究院有限公司 | Quick retrieval technology for face database |
| CN111009065A (en) * | 2019-12-09 | 2020-04-14 | 金现代信息产业股份有限公司 | Face recognition access control system optimization method and system based on clustering algorithm |
| CN111091106B (en) * | 2019-12-23 | 2023-10-10 | 浙江大华技术股份有限公司 | Image clustering method and device, storage medium and electronic device |
| CN111782855A (en) * | 2020-07-15 | 2020-10-16 | 上海依图网络科技有限公司 | Face image processing method, device, equipment and medium |
| CN112232290B (en) * | 2020-11-06 | 2023-12-08 | 四川云从天府人工智能科技有限公司 | Data clustering method, server, system and computer readable storage medium |
| CN112487222B (en) * | 2020-11-30 | 2021-11-30 | 江苏正赫通信息科技有限公司 | Method for quickly searching and effectively storing similar human faces |
| CN112559793B (en) * | 2021-02-23 | 2021-07-13 | 成都旺小宝科技有限公司 | Retrieval method of face image |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207898A (en) * | 2013-03-19 | 2013-07-17 | 天格科技（杭州）有限公司 | Method for rapidly retrieving similar faces based on locality sensitive hashing |
| CN104765768A (en) * | 2015-03-09 | 2015-07-08 | 深圳云天励飞技术有限公司 | Mass face database rapid and accurate retrieval method |
| CN106776849A (en) * | 2016-11-28 | 2017-05-31 | 西安交通大学 | A kind of method and guide system to scheme quick-searching sight spot |
-
2017
- 2017-08-30 CN CN201710760859.6A patent/CN110147455B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207898A (en) * | 2013-03-19 | 2013-07-17 | 天格科技（杭州）有限公司 | Method for rapidly retrieving similar faces based on locality sensitive hashing |
| CN104765768A (en) * | 2015-03-09 | 2015-07-08 | 深圳云天励飞技术有限公司 | Mass face database rapid and accurate retrieval method |
| CN106776849A (en) * | 2016-11-28 | 2017-05-31 | 西安交通大学 | A kind of method and guide system to scheme quick-searching sight spot |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110147455A (en) | 2019-08-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110147455B (en) | Face matching retrieval device and method | |
| US9804892B2 (en) | Tracking large numbers of moving objects in an event processing system | |
| CN107209853B (en) | Positioning and map construction method | |
| EP2742446B1 (en) | A system and method to store video fingerprints on distributed nodes in cloud systems | |
| US8706711B2 (en) | Descriptor storage and searches of k-dimensional trees | |
| US20150039538A1 (en) | Method for processing a large-scale data set, and associated apparatus | |
| US11100073B2 (en) | Method and system for data assignment in a distributed system | |
| CN105243060A (en) | Picture retrieval method and apparatus | |
| KR20130020050A (en) | Apparatus and method for managing bucket range of locality sensitivie hash | |
| Chen et al. | Metric similarity joins using MapReduce | |
| US10268655B2 (en) | Method, device, server and storage medium of searching a group based on social network | |
| US8370363B2 (en) | Hybrid neighborhood graph search for scalable visual indexing | |
| US11361195B2 (en) | Incremental update of a neighbor graph via an orthogonal transform based indexing | |
| WO2017072890A1 (en) | Data management system, data management method, and program | |
| CN110750671B (en) | Pedestrian retrieval method and device based on massive unstructured features | |
| CN110209895B (en) | Vector retrieval method, device and equipment | |
| US8666164B2 (en) | System and method for modeling a region segmented image | |
| US20220059195A1 (en) | Techniques for data-enabled drug discovery | |
| US20150142712A1 (en) | Rule discovery system, method, apparatus, and program | |
| CN113761017A (en) | Similarity searching method and device | |
| CN111639099A (en) | Full-text indexing method and system | |
| US10372917B1 (en) | Uniquely-represented B-trees | |
| US11620269B2 (en) | Method, electronic device, and computer program product for data indexing | |
| US20230100289A1 (en) | Searchable data processing operation documentation associated with data processing of raw data | |
| Tsapanos et al. | Large graph clustering using DCT-based graph clustering |
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 |