CN111709303A - Face image recognition method and device - Google Patents
Face image recognition method and device Download PDFInfo
- Publication number
- CN111709303A CN111709303A CN202010436932.6A CN202010436932A CN111709303A CN 111709303 A CN111709303 A CN 111709303A CN 202010436932 A CN202010436932 A CN 202010436932A CN 111709303 A CN111709303 A CN 111709303A
- Authority
- CN
- China
- Prior art keywords
- face image
- face
- image
- retrieved
- images
- 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.)
- Withdrawn
Links
Images
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
- 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
- G06F16/51—Indexing; Data structures therefor; Storage structures
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Library & Information Science (AREA)
- Databases & Information Systems (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Collating Specific Patterns (AREA)
Abstract
The embodiment of the application discloses a method and a device for recognizing a face image. The method comprises the following steps: after receiving a face image to be retrieved, determining a reference feature vector of the face image to be retrieved by using a preset face feature extraction model; comparing the reference characteristic vector with the characteristic vectors of all face images in a preset database to obtain candidate target face images corresponding to the face images; and selecting one face image from the candidate target face image and the face image as a target face image to serve as a recognition result of the face image to be retrieved.
Description
Technical Field
The embodiment of the application relates to the field of information processing, in particular to a method and a device for recognizing a face image.
Background
With the continuous development of the informatization technology, the biological characteristic information identification technology is widely applied in different fields without industries, and each person has unique biological characteristics, so that the biological characteristic identification technology becomes a more accurate, safer and more convenient identity authentication mode. Face recognition technology has become one of the important information recognition methods.
Face recognition belongs to an application field of image recognition. The traditional image recognition technology is mainly based on shallow information such as edge, shape, texture, color or feature point matching of an image, for example, methods such as HOG (Histogram of oriented Gradient), SIFT (Scale-invariant feature transform) and the like, the methods only extract the shallow features of the image, and when conditions such as image angle, definition, illumination, rotation and the like are obviously changed, the recognition accuracy is greatly reduced.
When searching and retrieving of specific people are carried out in the public security field, the text is used for retrieval in the related technology, so that the utilization rate of a large amount of image data collected from different sources is low, and the value of the image data cannot be fully mined and utilized.
Disclosure of Invention
In order to solve any technical problem, the embodiment of the application provides a method and a device for recognizing a face image.
In order to achieve the purpose of the embodiment of the present application, an embodiment of the present application provides a method for recognizing a face image, including:
after receiving a face image to be retrieved, determining a reference feature vector of the face image to be retrieved by using a preset face feature extraction model;
comparing the reference characteristic vector with the characteristic vectors of all face images in a preset database to obtain candidate target face images corresponding to the face images;
and selecting one face image from the candidate target face image and the face image as a target face image to serve as a recognition result of the face image to be retrieved.
An apparatus for recognizing a face image, comprising:
the determining module is used for determining a reference feature vector of the human face image to be retrieved by utilizing a preset human face feature extraction model after the human face image to be retrieved is received;
the comparison module is used for comparing the reference characteristic vector with the characteristic vectors of all face images in a preset database to obtain candidate target face images corresponding to the face images;
and the selection module is used for selecting one face image from the candidate target face image and the face image as a target face image as a recognition result of the face image to be retrieved.
A storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method as described above when executed.
An electronic device comprising a memory having a computer program stored therein and a processor arranged to execute the computer program to perform the method as described above.
One of the above technical solutions has the following advantages or beneficial effects:
after receiving a face image to be retrieved, determining a reference feature vector of the face image to be retrieved by using a preset face feature extraction model, comparing the reference feature vector with feature vectors of face images in a preset database to obtain a candidate target face image corresponding to the face image, selecting a face image from the candidate target face image and the face image as a target face image, using the face image as a recognition result of the face image to be retrieved, and comparing contents by using the feature vector of the face image to improve the comparison accuracy of the face features.
Additional features and advantages of the embodiments of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the embodiments of the application. The objectives and other advantages of the embodiments of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings are included to provide a further understanding of the embodiments of the present application and are incorporated in and constitute a part of this specification, illustrate embodiments of the present application and together with the examples of the embodiments of the present application do not constitute a limitation of the embodiments of the present application.
Fig. 1 is a flowchart of a method for recognizing a face image according to an embodiment of the present application;
fig. 2 is a structural diagram of a face recognition retrieval system based on a deep learning technique according to an embodiment of the present application;
fig. 3 is a structural diagram of a face image recognition apparatus according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application more apparent, the embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be noted that, in the embodiments of the present application, features in the embodiments and the examples may be arbitrarily combined with each other without conflict.
Fig. 1 is a flowchart of a face image recognition method according to an embodiment of the present application. As shown in fig. 1, the method shown in fig. 1 includes:
in an exemplary embodiment, the feature vector extraction may be performed by using a deep learning (Convolutional Neural Networks) model, wherein, when designing a specific network structure, the CNN model may be referred to or improved, for example, a model such as AlexNet, VGG, google lenet, ResNet, etc., the feature vector dimension may be 1024 dimensions, 512 dimensions, 256 dimensions, etc., and the specific dimension may be determined by the size of the data volume and the accuracy of retrieval.
in an exemplary embodiment, the reference feature vector is compared with the feature vector of the picture in the database to be queried, and N most similar pictures are queried, where N is an integer greater than or equal to 2.
And 103, selecting a face image from the candidate target face image and the face image as a target face image as a recognition result of the face image to be retrieved.
The method provided by the embodiment of the application utilizes the extraction model of the preset human face features to determine the reference feature vector of the human face image to be retrieved after receiving the human face image to be retrieved, compares the reference feature vector with the feature vector of each human face image in the preset database to obtain the candidate target human face image corresponding to the human face image, and selects one human face image as the target human face image from the candidate target human face image to serve as the identification result of the human face image to be retrieved, and performs content comparison through the feature vector of the human face image to improve the comparison accuracy of the human face features.
The method provided by the embodiments of the present application is explained as follows:
in an exemplary embodiment, before determining the feature vector of the facial image by using the preset extraction model of the facial features, the method further includes:
determining the position information of the face in the face image;
and removing other areas except the position information in the face image to obtain face data, and taking the face data as a processing object of a feature extraction model.
The position information may be labeled as coordinate information in the image;
detecting the region coordinates of a face part in an image uploaded by a user side, cutting the image according to the region coordinates, removing the part irrelevant to the face in the image, only reserving the region of the face part in the image, removing factors of other pixels in the image, and improving the accuracy of face feature extraction operation so as to further improve the retrieval accuracy, wherein a model for detecting the face region can adopt a target detection model for deep learning such as MTCNN, YOLO, SSD and the like;
in addition, the method can also comprise the steps of filtering the confidence coefficient threshold value of the detected face image, abandoning the face image with the confidence coefficient lower than the confidence coefficient threshold value, not taking the face image as the object to be retrieved, and taking the face image with the confidence coefficient greater than or equal to the confidence coefficient threshold value as the object to be retrieved.
In an exemplary embodiment, before comparing the reference feature vector with the feature vectors of the respective face images in the preset database, the method further includes:
constructing an index for the characteristic vector in the database to obtain an index table of the characteristic vector;
the step of comparing the feature vectors with the feature vectors of each face image in a preset database to obtain candidate target face images corresponding to the face images comprises the following steps:
determining a reference index of the reference feature vector in an index table of the feature vector;
and comparing the indexes of the feature vectors of the face images in the database by using the reference index to obtain a candidate target face image corresponding to the face image.
The characteristic vectors in the database can be indexed by using tools such as faiss, flann, annoy, kgraphh and the like, and different indexing modes can be selected according to the size of data volume.
By constructing the index of the feature vector, linear search can be avoided during vector comparison query, and efficiency vector index construction of comparison query is provided; especially when the data size is large, the query speed can be obviously improved by constructing the query index.
In an exemplary embodiment, the selecting a face image from the candidate target face image and the face image as a target face image as a recognition result of the face image to be retrieved includes:
calculating the picture similarity of each candidate target face image;
selecting candidate target face images with the image similarity larger than a preset similarity threshold value as face images to be selected according to the image similarity of each candidate target face image;
and selecting at least one face image from the face images to be selected as a target face image.
The similarity is calculated as the similarity between the inquired image and the image uploaded by the current user, and the similarity can be calculated by using Euclidean distance, cosine distance and the like.
Images with low similarity in the retrieved candidate target face images can be filtered out through the preset target similarity threshold, and the target face images can be conveniently selected subsequently.
In an exemplary embodiment, the selecting at least one face image from the face images to be selected as the target face image includes:
acquiring the description information of the identity of each person in each face image to be selected;
outputting the description information of the identity of the person in each face image to be selected;
and receiving a selection result of the face image to be selected according to the output description information of the identity of the person to obtain the target face image.
And further returning other auxiliary information of the corresponding person of the image, such as identity card number, name, age, household registration and other information according to the inquired image, so that the target face image can be conveniently selected according to the input other auxiliary information, and the accuracy of the identification result is improved.
In an exemplary embodiment, the selecting at least one face image from the face images to be selected as the target face image includes:
receiving a screening condition, wherein the screening condition comprises the identity characteristic information of people in the face image to be retrieved; acquiring the description information of the identity of the person in each face image to be selected;
and matching the description information of the identity of the person in each face image to be selected by using the identity characteristic information of the person in the face image to be retrieved to obtain the target face image.
The image is further filtered according to other auxiliary information of a person corresponding to the image, and the field information of the filtering can be specified by a user when the user inputs the detection image, such as identity card number, name, age, household registration and other information, so that the operation flexibility of the user is improved, and the accuracy of the identification result is improved.
In an exemplary embodiment, before the comparing the reference feature vector with the feature vector of each face image in a preset database to obtain a candidate target face image corresponding to the face image, the method further includes:
configuring a corresponding application identifier for the face image in the database, wherein the application identifier is used for recording business application scene information of the face image;
after receiving a face image to be retrieved, acquiring a business application scene corresponding to the face image to be retrieved;
searching the application identification of the face image in the database by using the service application scene corresponding to the face image to be retrieved to obtain the face image conforming to the service application scene;
and matching the face images in the face images conforming to the service application scene.
In the process of performing face region detection, feature vector extraction and feature vector index construction on a database, a unique identification field can be generated for marking the face image, and the field is used for marking the business application scene of the face image, so that a user side under different scenes can only search in the corresponding image retrieval library when the user side is used, the uniform management of the whole retrieval system is facilitated, and a plurality of sets of retrieval system services aiming at different image databases do not need to be deployed in a specific place.
The method provided by the embodiments of the present application is explained as follows:
the embodiment of the application provides a face recognition retrieval system based on a deep learning technology, and aims to provide more possibilities for information retrieval means and further fully mine and utilize the value of face image data.
Fig. 2 is a structural diagram of a face recognition retrieval system based on a deep learning technique according to an embodiment of the present application. As shown in fig. 2, the system includes:
the query image 01 is an image (i.e., a face image) which is requested by the user side to be retrieved and uploaded.
The face region detection 02 is to detect the coordinates of a face part in an image uploaded by a user side, cut the image according to region coordinates before face feature extraction 03, remove a part irrelevant to a face in the image, only reserve a region of the face part in the image, and improve the accuracy of the face feature extraction 03, so that the retrieval accuracy is further improved, the face region detection 02 model can be a deep learning target detection model such as MTCNN, YOLO, SSD, and the like, and the face region detection 02 comprises confidence threshold filtering of a detected face image.
The face feature extraction 03 is to extract feature vectors from a face image part subjected to target detection and clipping, the face feature extraction is the core of a face recognition retrieval system, the feature extraction adopts a CNN (convolutional neural network) model based on deep learning, wherein when a specific network structure is designed, a classic CNN model can be referred to or improved, for example, models such as AlexNet, VGG, google lenet, ResNet and the like, the feature vector dimensions can be 1024 dimensions, 512 dimensions, 256 dimensions and the like, the specific dimension selection needs to consider factors such as the size of a data volume and the retrieval accuracy, and the selection can be performed from low dimension to high dimension, the low dimension is preferentially selected, and when the effect is not good, the high dimension can be tried step by step.
The image database 04 is a face image library to be searched and compared, i.e. a search is performed in a designated face library.
The batch human face region detection 05 is to perform human face target detection on all pictures in the retrieved human face image library, the target detection model and the human face region detection 02 are the same model, wherein the batch human face region detection 05 comprises the steps of performing confidence threshold filtering on detected human face images and storing detection results in a database.
The face feature library 06 is a model which is obtained by converting all the pictures in the face image library into feature vectors after face target detection, and the extracted feature model and the face feature extraction 03 are the same model, wherein the extracted feature model and the face feature extraction 03 comprise the step of storing the feature vectors into a database.
The feature vector index 07 is an index constructed by feature vectors in the face feature library 06, so that linear search during vector comparison query is avoided, the vector index construction can be constructed by means of faiss and other tools, and different index modes can be selected according to the data size; especially when the data size is large, the query speed can be obviously improved by constructing the query index.
The feature vector retrieval 08 is to compare the feature vector extracted from the request picture uploaded by the user side with the feature vector of the picture in the face library to be queried corresponding to the request picture, and query the most similar picture of topN.
The similarity calculation 09 quantizes the similarity between the picture feature vector requested by the user side and the retrieved topN picture feature vector, and the similarity can be calculated by using the euclidean distance, the cosine distance and the like, wherein the smaller the computed euclidean distance or the smaller the cosine distance, the greater the similarity between the two pictures is.
The recall result sorting and filtering 10 is mainly based on filtering of a preset similarity threshold, so as to filter recall pictures with lower similarity, wherein a set value of the similarity threshold is determined according to the combination of factors such as model effect, service scene and the like.
The customized data content output 11 mainly returns other information of the corresponding person, such as identity card number, name, age, household registration and other information, according to the inquired image, wherein the information can be related to other databases according to the inquired image to inquire other required information; or filtered again according to some value, e.g. the user may specify attributes of household, age, etc. at the time of the query.
The synchronous update 12 is to update the related parts downstream in time when the image database 04 has an update action, including the update of the batch face region detection 05, the face feature library 06, and the feature vector index 07.
The CNN-based face target detection model and the feature extraction model are separately deployed as services during deployment, so that the purpose of model service decoupling is achieved, and later-stage model maintenance and updating are facilitated.
When the human face target detection model detects a plurality of human faces in the image, the human face part is intercepted in the later process, and only the region with the highest confidence coefficient is preferentially selected.
Aiming at image databases 04 of different sources or different scenes, in the process of batch human face region detection 05, human face feature library 06 and feature vector index 07, unique identification field carrying or configuration information is optionally generated, and when the user terminal under different scenes is used, the user terminal is only searched in the corresponding image retrieval library, so that the unified management of the whole retrieval system is facilitated, and a plurality of sets of retrieval system services aiming at different image databases do not need to be deployed in a specific place.
When the image database 04 has change operations such as deletion, addition and the like, the corresponding face feature library 06 should also have the same update operation and reconstruction of the feature vector index 07. The operation of the synchronous update 12 can be a timed task or real-time update, the timed period can be one hour, one day, one week and the like, and the specific update frequency is determined according to the requirements of the service scene; the real-time updating is to detect the state of the image database 04 in real time, and when a change operation occurs, the batch face region detection 05, the face feature library 06 and the feature vector index 07 are updated in time. The image database 04 is a face image library to be searched and compared, i.e. a search is performed in a designated face library.
The system provided by the embodiment of the application has the following advantages:
1. in the face detection before the face feature extraction, the features of only the face part in the image can be extracted through advanced face region detection, so that the interference of other parts of the image is avoided, and the retrieval accuracy is greatly improved;
2. compared with the extraction method in the related technology, the characteristic extraction method based on the CNN model has obvious effect improvement and can extract deep information; the construction of the feature vector index has obviously higher retrieval speed than linear search under large data volume, and the user experience is improved;
3. images with low similarity in the retrieved images can be filtered out through a preset target similarity threshold, and the recall accuracy is improved;
4. the other identity related information of the person corresponding to the image can be further returned according to the retrieved image, or further filtering can be performed according to the other identity related information, and the filtering field information can be specified by the user when the user inputs the detection image, so that the operation flexibility of the user is improved. Fig. 3 is a structural diagram of a face image recognition apparatus according to an embodiment of the present application. As shown in fig. 3, the apparatus shown in fig. 3 includes:
the first determination module is used for determining a reference feature vector of the face image to be retrieved by utilizing a preset face feature extraction model after the face image to be retrieved is received;
the comparison module is used for comparing the reference characteristic vector with the characteristic vectors of all face images in a preset database to obtain candidate target face images corresponding to the face images;
and the selection module is used for selecting one face image from the candidate target face image and the face image as a target face image as a recognition result of the face image to be retrieved.
In one exemplary embodiment, the apparatus further comprises:
the second determining module is used for determining the position information of the face in the face image before determining the feature vector of the face image;
and the processing module is used for removing other areas except the position information in the face image to obtain face data, and the face data is used as a processing object of the feature extraction model.
In one exemplary embodiment, the apparatus further comprises:
the construction module is used for constructing indexes for the characteristic vectors in the database to obtain an index table of the characteristic vectors before comparing the reference characteristic vectors with the characteristic vectors of the face images in the preset database;
the comparison module comprises:
a determining unit configured to determine a reference index of the reference feature vector in an index table of the feature vector;
and the comparison unit is arranged for comparing the indexes of the characteristic vectors of the face images in the database by using the reference index to obtain a candidate target face image corresponding to the face image.
In one exemplary embodiment, the selection module includes:
a calculation unit configured to calculate a picture similarity of each candidate target face image;
the first selection unit is arranged to select the candidate target face image with the image similarity larger than a preset similarity threshold value as the face image to be selected according to the image similarity of each candidate target face image;
and the second selection unit is used for selecting at least one face image from the face images to be selected as a target face image.
In one exemplary embodiment, the second selection unit includes:
the first acquisition subunit is configured to acquire the description information of the identity of each person in the face image to be selected;
the output subunit is configured to output description information of the identity of the person in each face image to be selected;
and the first receiving subunit is configured to receive a selection result of the face image to be selected according to the output description information of the identity of the person, so as to obtain the target face image.
In one exemplary embodiment, the second selection unit includes:
the second receiving subunit is configured to receive a screening condition, where the screening condition includes identity feature information of a person in the face image to be retrieved;
the second acquisition subunit is configured to acquire the description information of the identity of the person in each face image to be selected;
and the matching subunit is configured to match the description information of the identity of the person in each face image to be selected by using the identity characteristic information of the person in the face image to be retrieved to obtain the target face image.
In one exemplary embodiment, the apparatus further comprises:
the configuration module is configured to configure a corresponding application identifier for the face image in the database before comparing the reference feature vector with the feature vector of each face image in the preset database, wherein the application identifier is used for recording service application scene information of the face image;
the system comprises an acquisition module, a search module and a search module, wherein the acquisition module is used for acquiring a business application scene corresponding to a face image to be searched after the face image to be searched is received;
the searching module is arranged to search the application identification of the face image in the database by using the service application scene corresponding to the face image to be retrieved to obtain the face image conforming to the service application scene;
and the comparison module is set to match the face image in the face image conforming to the service application scene.
The device that this application embodiment provided, after receiving the face image that needs the retrieval, utilize the extraction model of the face characteristic of preset, confirm the benchmark eigenvector of the face image that needs the retrieval will benchmark eigenvector contrasts with the eigenvector of each face image in the database that sets up in advance, obtains the candidate target face image that face image corresponds, follows candidate target face image with face image selects a face image as target face image, as the identification result of the face image that needs the retrieval carries out the contrast of content through the eigenvector that utilizes face image, improves the comparison accuracy of face characteristic.
A storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method of any of the above when executed.
An electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to perform the method of any of the above.
It will be understood by those of ordinary skill in the art that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. In a hardware implementation, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be performed by several physical components in cooperation. Some or all of the components may be implemented as software executed by a processor, such as a digital signal processor or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as is well known to those of ordinary skill in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can accessed by a computer. In addition, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media as known to those skilled in the art.
Claims (10)
1. A face image recognition method comprises the following steps:
after receiving a face image to be retrieved, determining a reference feature vector of the face image to be retrieved by using a preset face feature extraction model;
comparing the reference characteristic vector with the characteristic vectors of all face images in a preset database to obtain candidate target face images corresponding to the face images;
and selecting one face image from the candidate target face image and the face image as a target face image to serve as a recognition result of the face image to be retrieved.
2. The method according to claim 1, wherein before determining the feature vector of the face image by using the preset extraction model of the face features, the method further comprises:
determining the position information of the face in the face image;
and removing other areas except the position information in the face image to obtain face data, and taking the face data as a processing object of a feature extraction model.
3. The method according to claim 1 or 2, characterized in that:
before comparing the reference feature vector with the feature vectors of the face images in the preset database, the method further includes:
constructing an index for the characteristic vector in the database to obtain an index table of the characteristic vector;
the step of comparing the feature vectors with the feature vectors of each face image in a preset database to obtain candidate target face images corresponding to the face images comprises the following steps:
determining a reference index of the reference feature vector in an index table of the feature vector;
and comparing the indexes of the feature vectors of the face images in the database by using the reference index to obtain a candidate target face image corresponding to the face image.
4. The method according to claim 1, wherein the selecting one face image from the candidate target face image and the face image as a target face image as a recognition result of the face image to be retrieved comprises:
calculating the picture similarity of each candidate target face image;
selecting candidate target face images with the image similarity larger than a preset similarity threshold value as face images to be selected according to the image similarity of each candidate target face image;
and selecting at least one face image from the face images to be selected as a target face image.
5. The method according to claim 4, wherein the selecting at least one face image from the face images to be selected as the target face image comprises:
acquiring the description information of the identity of each person in each face image to be selected;
outputting the description information of the identity of the person in each face image to be selected;
and receiving a selection result of the face image to be selected according to the output description information of the identity of the person to obtain the target face image.
6. The method according to claim 4, wherein the selecting at least one face image from the face images to be selected as the target face image comprises:
receiving a screening condition, wherein the screening condition comprises the identity characteristic information of people in the face image to be retrieved; acquiring the description information of the identity of the person in each face image to be selected;
and matching the description information of the identity of the person in each face image to be selected by using the identity characteristic information of the person in the face image to be retrieved to obtain the target face image.
7. The method according to claim 1, wherein before comparing the reference feature vector with the feature vectors of the face images in a preset database to obtain the candidate target face image corresponding to the face image, the method further comprises:
configuring a corresponding application identifier for the face image in the database, wherein the application identifier is used for recording business application scene information of the face image;
after receiving a face image to be retrieved, acquiring a business application scene corresponding to the face image to be retrieved;
searching the application identification of the face image in the database by using the service application scene corresponding to the face image to be retrieved to obtain the face image conforming to the service application scene;
and matching the face images in the face images conforming to the service application scene.
8. An apparatus for recognizing a face image, comprising:
the determining module is used for determining a reference feature vector of the human face image to be retrieved by utilizing a preset human face feature extraction model after the human face image to be retrieved is received;
the comparison module is used for comparing the reference characteristic vector with the characteristic vectors of all face images in a preset database to obtain candidate target face images corresponding to the face images;
and the selection module is used for selecting one face image from the candidate target face image and the face image as a target face image as a recognition result of the face image to be retrieved.
9. A storage medium, in which a computer program is stored, wherein the computer program is arranged to perform the method of any of claims 1 to 7 when executed.
10. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, and wherein the processor is arranged to execute the computer program to perform the method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010436932.6A CN111709303A (en) | 2020-05-21 | 2020-05-21 | Face image recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010436932.6A CN111709303A (en) | 2020-05-21 | 2020-05-21 | Face image recognition method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111709303A true CN111709303A (en) | 2020-09-25 |
Family
ID=72538019
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010436932.6A Withdrawn CN111709303A (en) | 2020-05-21 | 2020-05-21 | Face image recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111709303A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112364829A (en) * | 2020-11-30 | 2021-02-12 | 北京有竹居网络技术有限公司 | Face recognition method, device, equipment and storage medium |
| CN112699810A (en) * | 2020-12-31 | 2021-04-23 | 中国电子科技集团公司信息科学研究院 | Method and device for improving figure identification precision of indoor monitoring system |
| CN112836077A (en) * | 2021-01-21 | 2021-05-25 | 携程旅游网络技术（上海）有限公司 | Method, system, device and storage medium for searching specific person |
| CN114036334A (en) * | 2021-10-09 | 2022-02-11 | 武汉烽火信息集成技术有限公司 | Block chain-based face retrieval method, equipment and computer-readable storage medium |
| CN115495603A (en) * | 2022-09-26 | 2022-12-20 | 江苏衫数科技集团有限公司 | Clothing image retrieval method and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108932321A (en) * | 2018-06-29 | 2018-12-04 | 金蝶软件（中国）有限公司 | Research on face image retrieval, device, computer equipment and storage medium |
| WO2019071664A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Human face recognition method and apparatus combined with depth information, and storage medium |
| WO2019120115A1 (en) * | 2017-12-18 | 2019-06-27 | 深圳励飞科技有限公司 | Facial recognition method, apparatus, and computer apparatus |
| CN110297934A (en) * | 2019-07-04 | 2019-10-01 | 腾讯科技（深圳）有限公司 | A kind of image processing method, device and storage medium |
| CN110489951A (en) * | 2019-07-08 | 2019-11-22 | 招联消费金融有限公司 | Method, apparatus, computer equipment and the storage medium of risk identification |
-
2020
- 2020-05-21 CN CN202010436932.6A patent/CN111709303A/en not_active Withdrawn
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019071664A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Human face recognition method and apparatus combined with depth information, and storage medium |
| WO2019120115A1 (en) * | 2017-12-18 | 2019-06-27 | 深圳励飞科技有限公司 | Facial recognition method, apparatus, and computer apparatus |
| CN108932321A (en) * | 2018-06-29 | 2018-12-04 | 金蝶软件（中国）有限公司 | Research on face image retrieval, device, computer equipment and storage medium |
| CN110297934A (en) * | 2019-07-04 | 2019-10-01 | 腾讯科技（深圳）有限公司 | A kind of image processing method, device and storage medium |
| CN110489951A (en) * | 2019-07-08 | 2019-11-22 | 招联消费金融有限公司 | Method, apparatus, computer equipment and the storage medium of risk identification |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112364829A (en) * | 2020-11-30 | 2021-02-12 | 北京有竹居网络技术有限公司 | Face recognition method, device, equipment and storage medium |
| CN112699810A (en) * | 2020-12-31 | 2021-04-23 | 中国电子科技集团公司信息科学研究院 | Method and device for improving figure identification precision of indoor monitoring system |
| CN112699810B (en) * | 2020-12-31 | 2024-04-09 | 中国电子科技集团公司信息科学研究院 | Method and device for improving character recognition precision of indoor monitoring system |
| CN112836077A (en) * | 2021-01-21 | 2021-05-25 | 携程旅游网络技术（上海）有限公司 | Method, system, device and storage medium for searching specific person |
| CN112836077B (en) * | 2021-01-21 | 2024-03-12 | 携程旅游网络技术（上海）有限公司 | Specific person retrieval method, system, device and storage medium |
| CN114036334A (en) * | 2021-10-09 | 2022-02-11 | 武汉烽火信息集成技术有限公司 | Block chain-based face retrieval method, equipment and computer-readable storage medium |
| CN114036334B (en) * | 2021-10-09 | 2024-03-19 | 武汉烽火信息集成技术有限公司 | Face retrieval method, equipment and computer readable storage medium based on block chain |
| CN115495603A (en) * | 2022-09-26 | 2022-12-20 | 江苏衫数科技集团有限公司 | Clothing image retrieval method and system |
| CN115495603B (en) * | 2022-09-26 | 2023-11-24 | 江苏衫数科技集团有限公司 | Clothing image retrieval method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111709303A (en) | Face image recognition method and device | |
| CN106446816B (en) | Face recognition method and device | |
| EP3477506A1 (en) | Video detection method, server and storage medium | |
| US8270684B2 (en) | Automatic media sharing via shutter click | |
| CN103207898B (en) | A kind of similar face method for quickly retrieving based on local sensitivity Hash | |
| US11481432B2 (en) | Reverse image search method, apparatus and application system | |
| CN111125390B (en) | Database updating method and device, electronic equipment and computer storage medium | |
| CN110019891B (en) | Image storage method, image retrieval method and device | |
| US9665773B2 (en) | Searching for events by attendants | |
| CN113626444B (en) | Table query method, device, equipment and medium based on bitmap algorithm | |
| CN111368867B (en) | File classifying method and system and computer readable storage medium | |
| CN103428537A (en) | Video processing method and video processing device | |
| CN109800318A (en) | A kind of archiving method and device | |
| CN114937306A (en) | Target tracking method and system based on face clustering | |
| CN111309946B (en) | Established file optimization method and device | |
| CN107704471A (en) | A kind of information processing method and device and file call method and device | |
| CN105095468A (en) | Novel image retrieval method and system | |
| US11869127B2 (en) | Image manipulation method and apparatus | |
| CN112883213B (en) | Picture archiving method and device and electronic equipment | |
| CN115294368A (en) | Data screening method and device, electronic equipment and computer storage medium | |
| CN109213515A (en) | Normalizing method and device and an electronic equipment are buried under multi-platform | |
| WO2017143979A1 (en) | Image search method and device | |
| CN112069331B (en) | Data processing and searching method, device, equipment and storage medium | |
| CN114780780A (en) | Image retrieval method, image retrieval device, electronic equipment and storage medium | |
| CN111178409B (en) | Image matching and recognition system based on big data matrix stability analysis |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200925 |
|
| WW01 | Invention patent application withdrawn after publication |