CN112200084A - Face recognition method and device for video stream, electronic equipment and storage medium - Google Patents
Face recognition method and device for video stream, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112200084A CN112200084A CN202011079558.5A CN202011079558A CN112200084A CN 112200084 A CN112200084 A CN 112200084A CN 202011079558 A CN202011079558 A CN 202011079558A CN 112200084 A CN112200084 A CN 112200084A
- Authority
- CN
- China
- Prior art keywords
- face
- face image
- video frame
- video
- identity
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 230000001815 facial effect Effects 0.000 claims description 10
- 238000013507 mapping Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 4
- 238000012806 monitoring device Methods 0.000 description 11
- 238000000605 extraction Methods 0.000 description 9
- 238000001514 detection method Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 7
- 238000012544 monitoring process Methods 0.000 description 6
- 230000007547 defect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000010354 integration Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G06F16/58—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)
- Library & Information Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The embodiment of the invention discloses a face recognition method and device for video streaming, electronic equipment and a storage medium. The method comprises the following steps: acquiring a video stream; performing face recognition on the face image of each video frame in the video stream, and determining the identity corresponding to the face image of at least one video frame; carrying out face tracking on the face image of each video frame in the video stream, and endowing the same face identification to the face image which belongs to the same person and appears in each video frame in the video stream; and determining the identity corresponding to the face image of each video frame which is endowed with the same face identification in the video stream according to the identity corresponding to the face image of the at least one video frame and the endowed face identification. Based on the method and the device, the accuracy rate of face recognition of the video stream can be improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of computers, in particular to a face recognition method and device for video streaming, electronic equipment and a storage medium.
Background
The face recognition technology is a technology for recognizing a face by analyzing an input image or video stream based on a facial feature of a person. The method comprises the following steps: firstly, judging whether the input image has a face, if so, further giving the position and the size of each face and the position information of each main facial organ, then further extracting the characteristics contained in each face according to the information, and comparing the characteristics with the known face characteristics, thereby identifying the identity of each face.
However, the above-described prior art still has the following problems when applied to the non-perceptual face recognition.
The non-inductive face recognition is performed by non-artificial shooting, so that the situation that the face is seriously lateral and is partially shielded frequently occurs for continuous video streams. The computer can not accurately judge which frame of video frame has the best image quality and is most suitable for being sent to a feature extraction network for feature extraction, so that each frame of video frame in the video stream can only be sent to the feature extraction network for feature extraction. However, when a certain frame or a few frames in a video stream have a serious side face of a human face, a partial human face shielding condition, and the like, the difference between the human face feature information extracted from the video frames with poor quality and the known human face feature information is large, which further causes the false recognition of the human face in the video frames, and further causes various conditions to the recognition result of the continuous video frames of the same person appearing in the video stream, and the accuracy rate of the human face recognition is not high, which cannot meet the requirements of practical application.
Disclosure of Invention
It is an object of embodiments of the present invention to address at least the above problems and/or disadvantages and to provide at least the advantages described hereinafter.
The embodiment of the invention provides a face recognition method and device for a video stream, electronic equipment and a storage medium, which can improve the accuracy of face recognition in the video stream.
In a first aspect, a face recognition method for a video stream is provided, including:
acquiring a video stream;
performing face recognition on the face image of each video frame in the video stream, and determining the identity corresponding to the face image of at least one video frame;
carrying out face tracking on the face image of each video frame in the video stream, and endowing the same face identification to the face image which belongs to the same person and appears in each video frame in the video stream;
and determining the identity corresponding to the face image of each video frame which is endowed with the same face identification in the video stream according to the identity corresponding to the face image of the at least one video frame and the endowed face identification.
Optionally, the performing face recognition on the face image of each video frame in the video stream to determine the identity corresponding to the face image of at least one video frame includes:
matching the face image of each video frame in the video stream with a plurality of preset face images at different angles corresponding to each identity in a preset face image library;
counting the number of successfully matched preset face images corresponding to each identity in the preset face image library and the face images of each video frame in the video stream;
and if the face image of any frame of video frame in the video stream corresponds to a unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the unique identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity exceeds the preset number, determining the unique identity as the identity corresponding to the face image of the current video frame in the video stream.
Optionally, the performing face recognition on the face image of each video frame in the video stream to determine an identity corresponding to the face image of at least one video frame further includes:
if the facial image of any video frame in the video stream has at least two corresponding identities in the preset facial image library, the number of the successfully matched preset face images corresponding to the at least two identities is the largest among the number of the successfully matched preset face images corresponding to all the identities in the preset face image library, or the face image of any frame of video frame in the video stream has a unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the only one identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity is less than the preset number, then it is determined that face recognition of the face image of the current video frame in the video stream cannot be achieved.
Optionally, the matching the face image of each video frame in the video stream with a plurality of preset face images at different angles corresponding to each identity in a preset face image library includes:
extracting face characteristic information of face images of all video frames in the video stream;
calculating Euclidean distances between the face feature information extracted from the face image of each video frame in the video stream and the face feature information of each preset face image at different angles corresponding to each identity in the preset face image library;
the condition for successfully matching the face image of each video frame in the video stream with any one of the preset face images corresponding to each identity in the preset face image library is as follows:
and the Euclidean distance between the face characteristic information extracted from the face image of each video frame in the video stream and the face characteristic information of any preset face image corresponding to each identity in the preset face image library is smaller than a preset threshold value.
Optionally, the preset face images at the different angles are 5 face images at different angles, and the preset number is 2.
Optionally, the determining, according to the identity corresponding to the face image of the at least one video frame and the face identifier assigned thereto, the identity corresponding to the face image of the other video frame to which the face image of the at least one video frame is assigned the same face identifier in the video stream includes:
counting identities corresponding to face images endowed with the same face identification in the at least one frame of video frame in the video stream and the frequency of each identity recognized in the at least one frame of video frame;
and selecting the identity with the highest recognized frequency in the at least one frame of video frame as the identity corresponding to the face image of each video frame, which is endowed with the same face identification, in the video stream.
Optionally, the performing face recognition on the face image of each video frame in the video stream to determine an identity corresponding to the face image of at least one video frame further includes:
determining the position of the identified face image in the at least one video frame in the corresponding video frame;
the performing face tracking on the face image of each video frame in the video stream, and assigning the same face identifier to the face image belonging to the same person appearing in each video frame in the video stream, further includes:
determining the position of a face image endowed with a face identifier in each video frame in the video stream in the corresponding video frame;
before determining, according to the identity corresponding to the face image of the at least one video frame and the face identifier assigned thereto, the identity corresponding to the face image of each video frame to which the face image of the at least one video frame is assigned the same face identifier in the video stream, the method further includes:
matching the position of the face image with the recognized identity in the at least one frame of video frame in the corresponding video frame with the position of the face image with the face identification in the same frame of video frame;
and judging the face image with the recognized identity and the face image with the given face identification at the same position in the same frame of video frame as the same face image, and establishing a mapping relation between the identity corresponding to the same face image and the given face identification.
In a second aspect, a face recognition apparatus for video streaming is provided, including:
the acquisition module is used for acquiring a video stream;
the face recognition module is used for carrying out face recognition on the face image of each video frame in the video stream and determining the identity corresponding to the face image of at least one video frame;
the face tracking module is used for carrying out face tracking on the face image of each video frame in the video stream and endowing the same face identification to the face image which appears in each video frame in the video stream and belongs to the same person;
and the determining module is used for determining the identity corresponding to the face image of each video frame which is endowed with the same face identification in the video stream according to the identity corresponding to the face image of the at least one video frame and the endowed face identification.
Optionally, the face recognition module includes:
the matching unit is used for matching the face image of each video frame in the video stream with a plurality of preset face images in different angles corresponding to each identity in a preset face image library;
the statistical unit is used for counting the number of the face images of all the video frames in the video stream and the number of the successfully matched preset face images corresponding to all the identities in the preset face image library;
and the processing unit is used for determining the unique identity as the identity corresponding to the face image of the current video frame in the video stream if the face image of any frame of video frame in the video stream corresponds to the unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the unique identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity exceeds the preset number.
Optionally, the processing unit is further configured to:
if the facial image of any video frame in the video stream has at least two corresponding identities in the preset facial image library, the number of the successfully matched preset face images corresponding to the at least two identities is the largest among the number of the successfully matched preset face images corresponding to all the identities in the preset face image library, or the face image of any frame of video frame in the video stream has a unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the only one identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity is less than the preset number, then it is determined that face recognition of the face image of the current video frame in the video stream cannot be achieved.
Optionally, the matching unit is specifically configured to:
extracting face characteristic information of face images of all video frames in the video stream;
calculating Euclidean distances between the face feature information extracted from the face image of each video frame in the video stream and the face feature information of each preset face image at different angles corresponding to each identity in the preset face image library;
the condition for successfully matching the face image of each video frame in the video stream with any one of the preset face images corresponding to each identity in the preset face image library is as follows:
and the Euclidean distance between the face characteristic information extracted from the face image of each video frame in the video stream and the face characteristic information of any preset face image corresponding to each identity in the preset face image library is smaller than a preset threshold value.
Optionally, the preset face images at the different angles are 5 face images at different angles, and the preset number is 2.
Optionally, the determining module includes:
a counting unit, configured to count identities corresponding to face images that are given the same face identifier in the at least one video frame in the video stream and frequencies of the identities that are recognized in the at least one video frame;
and the selecting unit is used for selecting the identity with the highest identified frequency in the at least one frame of video frame as the identity corresponding to the face image of each video frame, which is endowed with the same face identification, in the video stream.
Optionally, the face recognition module is further specifically configured to:
determining the position of the identified face image in the at least one video frame in the corresponding video frame;
the face tracking module is further specifically configured to:
determining the position of a face image endowed with a face identifier in each video frame in the video stream in the corresponding video frame;
the device further comprises:
a matching module to:
matching the position of the face image with the recognized identity in the at least one frame of video frame in the corresponding video frame with the position of the face image with the face identification in the same frame of video frame;
and judging the face image with the recognized identity and the face image with the given face identification at the same position in the same frame of video frame as the same face image, and establishing a mapping relation between the identity corresponding to the same face image and the given face identification.
In a third aspect, an electronic device is provided, including: at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method described above.
In a fourth aspect, a storage medium is provided, on which a computer program is stored which, when executed by a processor, implements the method described above.
The embodiment of the invention at least comprises the following beneficial effects:
the method and the device for recognizing the face of the video stream provided by the embodiment of the invention firstly acquire the video stream, then perform face recognition on the face image of each video frame in the video stream, determine the identity corresponding to the face image of at least one video frame, then perform face tracking on the face image of each video frame in the video stream, endow the same face identification for the face image which belongs to the same person and appears in each video frame in the video stream, and finally determine the identity corresponding to the face image of each video frame which is endowed with the same face identification in the video stream according to the identity corresponding to the face image of at least one video frame and the endowed face identification. Based on the method and the device, the accuracy rate of face recognition of the video stream can be improved.
Additional advantages, objects, and features of embodiments of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of embodiments of the invention.
Drawings
Fig. 1 is a schematic diagram of an application scenario of a face recognition method for a video stream according to an embodiment of the present invention;
FIG. 2 is a flowchart of a face recognition method for a video stream according to an embodiment of the present invention;
FIG. 3 is a flowchart of a face recognition method for a single video frame according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a face recognition apparatus for video streaming according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The embodiments of the present invention will be described in further detail with reference to the accompanying drawings so that those skilled in the art can implement the embodiments of the invention with reference to the description.
Fig. 1 is a schematic view illustrating an application scenario of a face recognition method for a video stream according to an embodiment of the present invention. As shown in fig. 1, the face recognition method for video streaming can be applied to a monitoring system. In this application scenario, the monitoring system may include a monitoring device 110, a network 120, and a server 130, where the monitoring device 110 is connected to the server 130 through the network 120 for data interaction.
In particular, the monitoring device 110 is used to continuously capture images of a monitored area and generate a video stream. The monitoring device 110 may be a camera, video camera, or the like. The present invention is not particularly limited in this regard.
The server 130 may be a server providing various services, for example, a video stream sent from the monitoring device 110 is received through the network 120, the face images of the video frames in the video stream are subjected to face recognition through a data analysis capability, the identities corresponding to the face images of at least one video frame are determined, the face images of the video frames in the video stream are subjected to face tracking, the same face identifier is assigned to the face images of the same person appearing in the video frames in the video stream, and the identities corresponding to the face images of the video frames to which the same face identifier is assigned in the video stream are finally determined according to the identities corresponding to the face images of at least one video frame and the assigned face identifiers, so as to output a face recognition result for the video stream. The server 130 may be hardware or software. When the server 130 is hardware, it may be implemented as a distributed server cluster composed of a plurality of servers, or may be implemented as a single server. When the server is software, it can be implemented as a plurality of software or software modules, or as a single software or software module. The present invention is not particularly limited herein.
It should be noted that the video stream may be pre-stored in the server 130 by various means, or may be dynamically and remotely acquired from the monitoring device 110 in real time.
In addition, the above steps of data analysis and outputting the face recognition method for video stream, which are generally performed by the server, can be completely performed by the monitoring device 110 installed in the monitoring area, with the computing power meeting the requirements. Accordingly, a face recognition apparatus for video streaming may also be provided in the monitoring device 110. In this case, the monitoring system may not include the server 130 and the network 120.
It should be understood that the number of monitoring devices, networks and servers in fig. 1 is merely illustrative, and the number of monitoring devices, networks and servers may be selected according to actual needs. The present invention is not particularly limited in this regard.
Fig. 2 is a flowchart of a face recognition method for a video stream, which is performed by a system with processing capability, a server, or a face recognition device for a video stream according to an embodiment of the present invention. As shown in fig. 2, the method includes:
The acquired video stream is split, so that the video stream is split into a plurality of video frames. Here, a video frame is a minimum visual unit constituting a video stream, and each frame of the video frame is a still image. And performing operations such as face recognition, face tracking and the like on the split video frames.
In splitting the video stream into several video frames, the splitting of the video stream can be implemented by using various existing video framing methods or tools (e.g., Adobe premiere). The present invention is not particularly limited in this regard.
In some embodiments, each frame of video in the video stream is input into a face detection model for face detection, and whether a face exists in the current video frame is determined. The working process of the face detection model is generally as follows: and searching any one appointed image by adopting a certain strategy to determine whether a human face exists in the appointed image, and if so, returning the position, size, posture and the like of the human face detection frame. In this embodiment, an existing face detection model may be used, which is not specifically limited by the present invention.
Further, when a face is detected in the video frame, a face recognition model is adopted to perform face recognition on the detected face image. The working process of the face recognition model is generally as follows: and carrying out face recognition on the face image corresponding to the face detection frame generated by the face detection model to obtain a face recognition result corresponding to the face image. In this embodiment, an existing face recognition model (such as a feature extraction network) may be used, which is not specifically limited by the present invention.
In practical application, some video frames in a video stream may have situations such as a serious side face of a human face, a partial occlusion of the human face, and the like. Therefore, when performing face recognition on each video frame in the video stream in this step, only the identities corresponding to the face images of some of the video frames may be determined, and the identities corresponding to the face images of other video frames may not be recognized. In the subsequent operation, the video frames capable of identifying the corresponding identities of the face images are used as the basis to assist in determining the identities of the face images in other video frames, so as to realize the face identification of the video stream. In order to achieve the above purpose, it is necessary to ensure that the identity corresponding to the face image of at least one video frame in the video stream can be determined.
In the process of carrying out face detection on each frame of video frame in the video stream, if no face exists in the current video frame, the person entering the monitoring range can be considered to leave the monitoring range, and the face recognition result and the face tracking result of each video frame in a section of video stream before the current video frame can be subsequently processed and analyzed, so that the face recognition of the section of video stream is realized, and the face recognition result of the section of video stream is finally output. Here, "subsequent processing and analysis" includes processes of performing data integration on a face recognition result and a face tracking result of each video frame in a section of video stream before a current video frame, further establishing a mapping relationship between an identity corresponding to a face image in the video frame and an assigned face identifier, and determining an identity corresponding to a face image of each video frame, to which the same face identifier is assigned, in the video stream according to an identity corresponding to a face image of at least one video frame and an assigned face identifier.
In practical application, because the face recognition and the face tracking are performed separately, the computer does not know whether the face image with the recognized identity in a certain frame of video frame and the face image with the face identification belong to the same face image, that is, the computer cannot automatically link the face image, the identity corresponding to the face image, and the face identification given to the face image. However, for a single video frame, when a person appears in the video frame, the face image of the person necessarily appears in only one fixed position. Therefore, the relationship between the face image processed in the face recognition step and the face image processed in the face tracking step can be established by the position of the face image in the corresponding video frame.
That is, in some embodiments, when performing face recognition on each video frame in a video stream, for a face image of a certain frame of video frame, in addition to determining an identity corresponding to the face image, a position of the face image in the corresponding video frame is also determined. Specifically, the coordinates of the identified face image in the corresponding video frame can be used to represent the position of the face image in the corresponding video frame. Based on this, the face recognition result actually output in this step includes the identity (which may be represented by name) corresponding to the face image in the video frame and the coordinates of the face image with the identified identity in the corresponding video frame.
And step 230, performing face tracking on the face image of each video frame in the video stream, and assigning the same face identification to the face image which belongs to the same person and appears in each video frame in the video stream.
Specifically, a face tracking model is adopted to track the face of the face image of each video frame in the video stream. The face tracking model refers to a model for capturing and tracking a face in consecutive images. In this embodiment, the face tracking for each video frame may be implemented by using an existing face tracking model, which is not specifically limited in the present invention.
In some embodiments, the face identification may be specified by an id value. The face tracking model judges whether the face appearing in the current video frame and the face appearing in the previous video frame belong to the same person, if so, the same id value as the previous frame is returned, and if not, a new id value is returned. Based on the above processing of the face tracking model, the same face identification is finally given to the face images which appear in each video frame and belong to the same person, and correspondingly, different face identifications are given to the face images which appear in each video frame and do not belong to the same person.
In some embodiments, in order to establish a relationship between the face image processed in the face recognition step and the face image processed in the face tracking step in a certain frame of video frame, that is, in order to establish a mapping relationship between the face image in a certain frame of video frame, the identity corresponding to the face image, and the face identifier assigned to the face image, when the face image of each video frame in the video stream is subjected to face tracking, the positions of the face images assigned with the face identifiers in the video frames are determined, except that the same face identifier is assigned to the face images belonging to the same person appearing in each video frame in the video stream. Specifically, the position of the face image in the corresponding video frame may be represented by coordinates of the face image to which the face identification is given in the corresponding video frame. Based on this, the face tracking result actually output in this step includes the face identifier assigned to the face image in the video frame and the coordinates of the face image assigned with the face identifier in the corresponding video frame.
Further, the position of the face image with the recognized identity in at least one frame of video frame in the corresponding video frame is matched with the position of the face image with the face identification in the same frame of video frame. And judging the face image with the recognized identity and the face image with the given face identification at the same position in the same frame of video frame as the same face image, and establishing a mapping relation between the identity corresponding to the same face image and the given face identification.
Specifically, the coordinates of the face image with the recognized identity in the video frame are compared with the coordinates of the face image with the face identification, if the two coordinates are consistent, the face image with the recognized identity in the face recognition step and the face image with the face identification in the face tracking step in the video frame are the same face image, so that the face image, the identity corresponding to the face image and the face identification given to the face image can be linked, and the mapping relation between the identity corresponding to the face image and the face identification given to the face image is established. Through the above data integration process, for a single video frame, a set of data in the following form can be output: [ [ coordinates, name, id value ], …, [ coordinates, name, id value ], [ n ], … ], where n represents the number of face images whose identity is recognized in the current video frame. Then, the above-mentioned data integration process may be performed on all video frames containing the face images with the recognized identities, and assuming that there are m frames in such video frames, m pieces of data represented in the above-mentioned form may be finally output.
It should be noted that, for those video frames in which the face image cannot be recognized, although the face image also exists in the video frames, the face image therein is not recognized effectively, and thus the above-mentioned data integration processing cannot be performed on the video frames. Only face images and face identifications given to the face images are obtained for the video frames.
In practical application, some video frames in a video stream have a situation that a face is seriously detected or the face is blocked, and a face recognition result may not be obtained for face images in the video frames. The embodiment can identify the identity corresponding to the face image in the video frames.
Although the human face image of a person appears with an occlusion in some video frames, the identity of the human face image cannot be recognized. However, the identity of the face image of at least one video frame can still be determined by performing step 220.
Whereas for the same person appearing in the video stream he appears continuously in the video frames. By executing step 230, the face images belonging to the same person in each video frame in the video stream can be tracked, and the face images belonging to the same person are all given the same face identification. It should be understood that whether the face image in the current video frame has an occlusion or a severe side face, whether the face image is the same person as the face image in the previous video frame can be determined through face tracking, and then it is determined to assign a new face identifier to the face image or a face identifier the same as the face image in the previous video frame.
For a video frame containing a face image with a recognized identity, the face image is also endowed with face identification. Therefore, based on the face image in the same frame of video frame, a mapping relationship between the face image, the identity corresponding to the face image, and the face identifier assigned to the face image can be established, and more simply, the mapping relationship between the identity and the face identifier is established. Accordingly, the identities corresponding to those face images that are given the same face identification in other video frames of the video stream can be determined.
For example, it is assumed that in 20 frames of video frames, face recognition is performed on face images in 13 frames, the obtained face recognition result is a small sheet, and the identity of the face image cannot be recognized in 7 frames due to the defect of the face image. The same face identification is given to one face image in each frame of video frame through face tracking, and the id values are all 1, so that the same person appears in 20 frames of video frames, and the id value of the person is 1. Then, according to the face recognition result of the first 13 frames of video frames, it is known that the identity corresponding to the face image with id value 1 is a small sheet, and it can be further presumed that the identity corresponding to the face image with id value 1 in the other 7 frames of video frames is also a small sheet. That is, this analysis result may be used as a face recognition result for the video stream.
In practical applications, besides the situation that the identity cannot be recognized due to the defect that the side face of the face is serious or the face is blocked in the video frames, the false recognition may occur, that is, for the video frames, the identity of a face image can be obtained through face recognition, but the recognized identity may be wrong. The embodiment can also realize the discrimination of wrong face recognition results in the video stream, thereby improving the accuracy of the face recognition of the video stream.
In some embodiments, determining, according to the identity corresponding to the face image of the at least one video frame and the assigned face identifier, the identity corresponding to the face image of each video frame to which the face image of the at least one video frame is assigned the same face identifier in the video stream includes: counting identities corresponding to face images endowed with the same face identification in at least one frame of video frame in the video stream and the frequency of each identity recognized in at least one frame of video frame; and selecting the identity with the highest recognized frequency in at least one video frame as the identity corresponding to the face image of each video frame, which is endowed with the same face identification with the face image of at least one video frame, in the video stream.
For all video frames containing face images with recognized identities, assuming that the face identifiers given to the face images in the video frames are the same and are face identifiers C, counting the identities corresponding to the face images in the video frames, for example, if the identity corresponding to the face image in one part of the video frames is a and the identity corresponding to the face image in the other part of the video frames is B, continuously counting the recognized frequencies of the identity a and the identity B in all the video frames, wherein the recognized frequency of a certain identity in all the video frames is the ratio of the recognized times of the identity to all the video frames. It should be noted that "all video frames" herein refer to those video frames that contain face images that are given the same face identification and are recognized. When the face image of a certain video frame is not identified, the video frame obviously cannot be used for calculating the frequency of identifying a certain identity. When more than one recognized face image exists in a certain video frame, the identity corresponding to one face image is counted, and the number of times that the identity corresponding to the face image is recognized in other video frames is only considered.
In the above example, based on the statistical results of identity a and identity B, one of the identities with the highest recognition frequency is selected as the identity corresponding to the face image with the same face identification in each video frame in the video stream. Assuming that the identification frequency of the identity a is higher than the identification frequency of the identity B, it is considered that the identity a belongs to the correct face identification result and the identity B belongs to the wrong face identification result for the face image to which the face identifier C is assigned, and therefore the identity a is taken as the identity corresponding to the face image to which the face identifier C is assigned in the video stream.
Based on the above process, the present embodiment eliminates the situation of false recognition of some video frames in the video stream, and further improves the accuracy of face recognition on the video stream.
In a specific example, it is assumed that in 20 frames of video frames, face recognition is implemented on face images in 13 frames, wherein the face recognition result of 10 frames is a small page, the face recognition result of 3 frames is a king, and the identity of the face image cannot be recognized in 7 frames due to the defect of the face image. The same face identification is given to the face image in each frame of video frame through face tracking, and the id values are all 1, so that the same person appears in the 20 frames of video frames, the id value of the person is 1, and two identities of the Xiaozhang and the Xiaowang are respectively identified. Counting the recognized frequency of the two identities, wherein the recognized frequency of the identity Xiaozhang is 10/13, and the recognized frequency of the identity Xiaowang is 3/13, so that the identity Xiaozhang with the highest recognized frequency is judged to be a real name, and the identity Xiaowang is an error name. Finally, the identity "sheetlet" is taken as the identity corresponding to the face image with the id value of 1 in the 20 frames. That is, this analysis result may be used as a face recognition result for the video stream.
In some embodiments, the false recognition rate of the face image in a single video frame can be reduced by means of statistical judgment. Figure 3 shows a flow chart for face recognition for a single video frame. As shown in fig. 3, step 220, performing face recognition on the face image of each video frame in the video stream, and determining the identity corresponding to the face image of at least one video frame, further includes:
and 310, matching the face image of each video frame in the video stream with a plurality of preset face images at different angles corresponding to each identity in a preset face image library.
The preset face image library is used for storing preset face images so as to compare and analyze the preset face images with the face images in the video frames. And for each identity, a plurality of preset face images at different angles are respectively provided. In some examples, 5 preset face images at different angles may be provided, for example, a front face image, left and right side face images, and face images for looking up and down. The face images of the video frames are matched with the preset face images at different angles, so that the accuracy rate of recognizing the face lines in the video frames is improved.
In some examples, matching the face image of each video frame in the video stream with a plurality of preset face images at different angles corresponding to each identity in a preset face image library includes: extracting face characteristic information of a face image of each video frame in a video stream; calculating Euclidean distances between face feature information extracted from the face image of each video frame in the video stream and face feature information of each preset face image at different angles corresponding to each identity in a preset face image library; then, the successful matching of the face image of each video frame in the video stream and any one of the preset face images corresponding to each identity in the preset face image library is determined as follows: and the Euclidean distance between the face characteristic information extracted from the face image of each video frame in the video stream and the face characteristic information of any preset face image corresponding to each identity in the preset face image library is smaller than a preset threshold value. The face image of a single video frame can be input into a feature extraction network for feature extraction. Accordingly, a preset face image corresponding to a certain identity in the preset face image library can be input into the feature extraction network in advance for feature extraction, and the extracted face feature information can be stored in the preset face image library.
In a specific example, when 5 preset face images with different angles are adopted, 5 euclidean distances can be calculated for any identity in the preset face image library. For the convenience of statistics, a table with N rows and 5 columns may be established, where each row represents a person (specifically, the identity of the person may be represented by a name) stored in the preset face image library, and each column represents a euclidean distance between the face feature information extracted from the face image in the video frame and the face feature information of the preset face image at one angle. Then, a voting counting method in statistics is adopted at this time, each line in the table is traversed, and if a plurality of the 5 Euclidean distances in the current line are smaller than a preset threshold value, a plurality of votes are cast for the name corresponding to the current line. Here, if the euclidean distance is smaller than the preset threshold, it indicates that the face image in the video frame is successfully matched with one of the preset face images of a certain identity in the preset face image library.
And step 320, counting the number of the successfully matched preset face images corresponding to the identities in the preset face image library and the face images of the video frames in the video stream.
In a specific example, for convenience of statistics, a table with N rows and 1 column may be still established, where each row represents a person stored in the preset face image library, and each column represents the number of votes corresponding to the person, that is, the number of preset face images in the video frame, which are successfully matched with a certain identity in the preset face image library. When 5 preset face images with different angles are adopted, the voting number is a number between 0 and 5. By looking at the table, it can be determined whether the maximum number of votes and the known identity to which the maximum number of votes corresponds is unique.
Step 330, if the face image of any frame of video frame in the video stream corresponds to a unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the unique identity is the largest among the number of the successfully matched preset face images corresponding to all the identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity exceeds the preset number, determining the unique identity as the identity corresponding to the face image of the current video frame in the video stream.
Specifically, for the face image of a single video frame, the known identity corresponding to the maximum vote number is unique, and the maximum vote number exceeds the preset number, the face image of the video frame is considered to realize effective identification, and the corresponding name is returned.
In some examples, when 5 preset face images at different angles are used, the preset number is set to 2. Based on the setting, whether the face recognition of a single video frame is an effective face recognition result can be accurately judged, and the face recognition accuracy rate of the single video frame is further improved.
Further, if the face image of any video frame in the video stream corresponds to at least two identities in the preset face image library, the number of the successfully matched preset face images corresponding to the at least two identities is the largest in the number of the successfully matched preset face images corresponding to all the identities in the preset face image library at the same time, or the face image of any video frame in the video stream corresponds to a single identity in the preset face image library, the number of the successfully matched preset face images corresponding to the single identity is the largest in the number of the successfully matched preset face images corresponding to all the identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the single identity is smaller than the preset number, it is determined that face recognition of the face image of the current video frame in the video stream cannot be achieved.
Specifically, for the face image of a single video frame, the known identity corresponding to the maximum vote number is not unique, that is, there are two or more known identities, and the number of the preset face images successfully matched is the same and is the largest, in this case, it is considered that the face image in the video frame cannot be effectively identified, and "unknown" is returned. In addition, for the face image of a single video frame, the known identity corresponding to the maximum vote number is unique, but the maximum vote number does not exceed the preset number, the face image in the video frame cannot be effectively identified, and the method returns to 'unknown'.
To sum up, the face recognition method for video streams provided by the embodiments of the present invention first obtains a video stream, then performs face recognition on a face image of each video frame in the video stream, determines an identity corresponding to the face image of at least one video frame, then performs face tracking on the face image of each video frame in the video stream, assigns the same face identifier to a face image belonging to the same person appearing in each video frame in the video stream, and finally determines an identity corresponding to the face image of each video frame, to which the same face identifier is assigned, in the video stream according to the identity corresponding to the face image of at least one video frame and the face identifier assigned thereto. Based on the method, the problems that the identity cannot be effectively recognized or the identity cannot be recognized by mistake and the like caused by the defects that the human face of some video frames in the video stream is seriously side face or the human face is shielded and the like can be solved, and the accuracy of the human face recognition of the video stream is improved.
Fig. 4 is a schematic structural diagram illustrating a face recognition apparatus for video streaming according to an embodiment of the present invention. As shown in fig. 4, the face recognition apparatus 400 for video streaming includes: an obtaining module 410, configured to obtain a video stream; a face recognition module 420, configured to perform face recognition on a face image of each video frame in the video stream, and determine an identity corresponding to the face image of at least one video frame; a face tracking module 430, configured to perform face tracking on a face image of each video frame in the video stream, and assign the same face identifier to a face image that belongs to the same person and appears in each video frame in the video stream; the determining module 440 is configured to determine, according to the identity corresponding to the face image of the at least one video frame and the assigned face identifier, the identity corresponding to the face image of each video frame in the video stream to which the face image of the at least one video frame is assigned the same face identifier.
In some embodiments, the face recognition module comprises: the matching unit is used for matching the face image of each video frame in the video stream with a plurality of preset face images in different angles corresponding to each identity in a preset face image library; the statistical unit is used for counting the number of the face images of all the video frames in the video stream and the number of the successfully matched preset face images corresponding to all the identities in the preset face image library; and the processing unit is used for determining the unique identity as the identity corresponding to the face image of the current video frame in the video stream if the face image of any frame of video frame in the video stream corresponds to the unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the unique identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity exceeds the preset number.
In some embodiments, the processing unit is further configured to: if the facial image of any video frame in the video stream has at least two corresponding identities in the preset facial image library, the number of the successfully matched preset face images corresponding to the at least two identities is the largest among the number of the successfully matched preset face images corresponding to all the identities in the preset face image library, or the face image of any frame of video frame in the video stream has a unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the only one identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity is less than the preset number, then it is determined that face recognition of the face image of the current video frame in the video stream cannot be achieved.
In some embodiments, the matching unit is specifically configured to: extracting face characteristic information of face images of all video frames in the video stream; calculating Euclidean distances between the face feature information extracted from the face image of each video frame in the video stream and the face feature information of each preset face image at different angles corresponding to each identity in the preset face image library; the condition for successfully matching the face image of each video frame in the video stream with any one of the preset face images corresponding to each identity in the preset face image library is as follows: and the Euclidean distance between the face characteristic information extracted from the face image of each video frame in the video stream and the face characteristic information of any preset face image corresponding to each identity in the preset face image library is smaller than a preset threshold value.
In some embodiments, the plurality of preset face images at different angles are 5 face images at different angles, and the preset number is 2.
In some embodiments, the determining module comprises: a counting unit, configured to count identities corresponding to face images that are given the same face identifier in the at least one video frame in the video stream and frequencies of the identities that are recognized in the at least one video frame; and the selecting unit is used for selecting the identity with the highest identified frequency in the at least one frame of video frame as the identity corresponding to the face image of each video frame, which is endowed with the same face identification, in the video stream.
In some embodiments, the face recognition module is further specifically configured to: determining the position of the identified face image in the at least one video frame in the corresponding video frame; the face tracking module is further specifically configured to: determining the position of a face image endowed with a face identifier in each video frame in the video stream in the corresponding video frame; the device further comprises: a matching module to: matching the position of the face image with the recognized identity in the at least one frame of video frame in the corresponding video frame with the position of the face image with the face identification in the same frame of video frame; and judging the face image with the recognized identity and the face image with the given face identification at the same position in the same frame of video frame as the same face image, and establishing a mapping relation between the identity corresponding to the same face image and the given face identification.
Fig. 5 shows an electronic device of an embodiment of the invention. As shown in fig. 5, the electronic device 500 includes: at least one processor 510, and a memory 520 communicatively coupled to the at least one processor 510, wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method.
Specifically, the memory 520 and the processor 510 are connected together via a bus 530, and can be a general-purpose memory and a processor, which are not specifically limited herein, and when the processor 510 executes a computer program stored in the memory 520, the operations and functions described in the embodiments of the present invention in conjunction with fig. 1 to 4 can be performed.
An embodiment of the present invention further provides a storage medium, on which a computer program is stored, which, when executed by a processor, implements the method. For specific implementation, reference may be made to the method embodiment, which is not described herein again.
While embodiments of the present invention have been disclosed above, it is not limited to the applications listed in the description and the embodiments. It is fully applicable to a variety of fields in which embodiments of the present invention are suitable. Additional modifications will readily occur to those skilled in the art. Therefore, the embodiments of the invention are not to be limited to the specific details and illustrations shown and described herein, without departing from the general concept defined by the claims and their equivalents.
Claims (10)
1. A face recognition method for a video stream, comprising:
acquiring a video stream;
performing face recognition on the face image of each video frame in the video stream, and determining the identity corresponding to the face image of at least one video frame;
carrying out face tracking on the face image of each video frame in the video stream, and endowing the same face identification to the face image which belongs to the same person and appears in each video frame in the video stream;
and determining the identity corresponding to the face image of each video frame which is endowed with the same face identification in the video stream according to the identity corresponding to the face image of the at least one video frame and the endowed face identification.
2. The method as claimed in claim 1, wherein the performing face recognition on the face image of each video frame in the video stream to determine the identity corresponding to the face image of at least one video frame comprises:
matching the face image of each video frame in the video stream with a plurality of preset face images at different angles corresponding to each identity in a preset face image library;
counting the number of successfully matched preset face images corresponding to each identity in the preset face image library and the face images of each video frame in the video stream;
and if the face image of any frame of video frame in the video stream corresponds to a unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the unique identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity exceeds the preset number, determining the unique identity as the identity corresponding to the face image of the current video frame in the video stream.
3. The face recognition method for video streams according to claim 2, wherein the face recognition is performed on the face image of each video frame in the video stream to determine the identity corresponding to the face image of at least one video frame, further comprising:
if the facial image of any video frame in the video stream has at least two corresponding identities in the preset facial image library, the number of the successfully matched preset face images corresponding to the at least two identities is the largest among the number of the successfully matched preset face images corresponding to all the identities in the preset face image library, or the face image of any frame of video frame in the video stream has a unique identity in the preset face image library, the number of the successfully matched preset face images corresponding to the only one identity is the largest among the number of the successfully matched preset face images corresponding to all identities in the preset face image library, and the number of the successfully matched preset face images corresponding to the unique identity is less than the preset number, then it is determined that face recognition of the face image of the current video frame in the video stream cannot be achieved.
4. The face recognition method for video streaming according to claim 2 or 3, wherein the matching of the face image of each video frame in the video streaming with a plurality of preset face images at different angles corresponding to each identity in a preset face image library comprises:
extracting face characteristic information of face images of all video frames in the video stream;
calculating Euclidean distances between the face feature information extracted from the face image of each video frame in the video stream and the face feature information of each preset face image at different angles corresponding to each identity in the preset face image library;
the condition for successfully matching the face image of each video frame in the video stream with any one of the preset face images corresponding to each identity in the preset face image library is as follows:
and the Euclidean distance between the face characteristic information extracted from the face image of each video frame in the video stream and the face characteristic information of any preset face image corresponding to each identity in the preset face image library is smaller than a preset threshold value.
5. The method according to claim 3, wherein the plurality of preset face images at different angles are 5 face images at different angles, and the preset number is 2.
6. The method as claimed in claim 1, wherein the determining the identity of the face image of the video stream corresponding to the other video frame to which the face image of the at least one video frame is assigned the same face identification according to the identity of the face image of the at least one video frame and the assigned face identification comprises:
counting identities corresponding to face images endowed with the same face identification in the at least one frame of video frame in the video stream and the frequency of each identity recognized in the at least one frame of video frame;
and selecting the identity with the highest recognized frequency in the at least one frame of video frame as the identity corresponding to the face image of each video frame, which is endowed with the same face identification, in the video stream.
7. The face recognition method for a video stream as set forth in claim 1,
the performing face recognition on the face image of each video frame in the video stream to determine the identity corresponding to the face image of at least one video frame further includes:
determining the position of the identified face image in the at least one video frame in the corresponding video frame;
the performing face tracking on the face image of each video frame in the video stream, and assigning the same face identifier to the face image belonging to the same person appearing in each video frame in the video stream, further includes:
determining the position of a face image endowed with a face identifier in each video frame in the video stream in the corresponding video frame;
before determining, according to the identity corresponding to the face image of the at least one video frame and the face identifier assigned thereto, the identity corresponding to the face image of each video frame to which the face image of the at least one video frame is assigned the same face identifier in the video stream, the method further includes:
matching the position of the face image with the recognized identity in the at least one frame of video frame in the corresponding video frame with the position of the face image with the face identification in the same frame of video frame;
and judging the face image with the recognized identity and the face image with the given face identification at the same position in the same frame of video frame as the same face image, and establishing a mapping relation between the identity corresponding to the same face image and the given face identification.
8. A face recognition apparatus for use in video streaming, comprising:
the video stream acquisition module is used for acquiring a video stream;
the face recognition module is used for carrying out face recognition on the face image of each video frame in the video stream and determining the identity corresponding to the face image of at least one video frame;
the face tracking module is used for carrying out face tracking on the face image of each video frame in the video stream and endowing the same face identification to the face image which appears in each video frame in the video stream and belongs to the same person;
and the identity determining module is used for determining the identity corresponding to the face image of each video frame which is endowed with the same face identification in the video stream according to the identity corresponding to the face image of the at least one video frame and the endowed face identification.
9. An electronic device, comprising: at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method of any of claims 1-8.
10. A storage medium on which a computer program is stored, which program, when being executed by a processor, is adapted to carry out the method of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011079558.5A CN112200084A (en) | 2020-10-10 | 2020-10-10 | Face recognition method and device for video stream, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011079558.5A CN112200084A (en) | 2020-10-10 | 2020-10-10 | Face recognition method and device for video stream, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112200084A true CN112200084A (en) | 2021-01-08 |
Family
ID=74013319
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011079558.5A Pending CN112200084A (en) | 2020-10-10 | 2020-10-10 | Face recognition method and device for video stream, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112200084A (en) |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014075495A1 (en) * | 2012-11-16 | 2014-05-22 | 中兴通讯股份有限公司 | Face recognition tracking method and system |
| KR101549599B1 (en) * | 2014-06-02 | 2015-09-02 | 고권태 | Entrance Warning System of Restricted Areas Capable of Verification and Tracking Using Face Image Recognition and Tag Recognition |
| CN106845385A (en) * | 2017-01-17 | 2017-06-13 | 腾讯科技（上海）有限公司 | The method and apparatus of video frequency object tracking |
| CN107609497A (en) * | 2017-08-31 | 2018-01-19 | 武汉世纪金桥安全技术有限公司 | The real-time video face identification method and system of view-based access control model tracking technique |
| CN107644204A (en) * | 2017-09-12 | 2018-01-30 | 南京凌深信息科技有限公司 | A kind of human bioequivalence and tracking for safety-protection system |
| CN107679613A (en) * | 2017-09-30 | 2018-02-09 | 同观科技（深圳）有限公司 | A kind of statistical method of personal information, device, terminal device and storage medium |
| CN108171207A (en) * | 2018-01-17 | 2018-06-15 | 百度在线网络技术（北京）有限公司 | Face identification method and device based on video sequence |
| CN109815858A (en) * | 2019-01-10 | 2019-05-28 | 中国科学院软件研究所 | A kind of target user Gait Recognition system and method in surroundings |
| CN109919977A (en) * | 2019-02-26 | 2019-06-21 | 鹍骐科技（北京）股份有限公司 | A kind of video motion personage tracking and personal identification method based on temporal characteristics |
| CN110232323A (en) * | 2019-05-13 | 2019-09-13 | 特斯联（北京）科技有限公司 | A kind of parallel method for quickly identifying of plurality of human faces for crowd and its device |
| CN110503059A (en) * | 2019-08-27 | 2019-11-26 | 国网电子商务有限公司 | A kind of face identification method and system |
| CN110941978A (en) * | 2019-05-23 | 2020-03-31 | 罗普特科技集团股份有限公司 | Face clustering method and device for unidentified personnel and storage medium |
| CN111079670A (en) * | 2019-12-20 | 2020-04-28 | 北京百度网讯科技有限公司 | Face recognition method, face recognition device, face recognition terminal and face recognition medium |
| CN111310731A (en) * | 2019-11-15 | 2020-06-19 | 腾讯科技（深圳）有限公司 | Video recommendation method, device and equipment based on artificial intelligence and storage medium |
| CN111401171A (en) * | 2020-03-06 | 2020-07-10 | 咪咕文化科技有限公司 | Face image recognition method and device, electronic equipment and storage medium |
| CN111444817A (en) * | 2020-03-24 | 2020-07-24 | 咪咕文化科技有限公司 | Person image identification method and device, electronic equipment and storage medium |
-
2020
- 2020-10-10 CN CN202011079558.5A patent/CN112200084A/en active Pending
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014075495A1 (en) * | 2012-11-16 | 2014-05-22 | 中兴通讯股份有限公司 | Face recognition tracking method and system |
| KR101549599B1 (en) * | 2014-06-02 | 2015-09-02 | 고권태 | Entrance Warning System of Restricted Areas Capable of Verification and Tracking Using Face Image Recognition and Tag Recognition |
| CN106845385A (en) * | 2017-01-17 | 2017-06-13 | 腾讯科技（上海）有限公司 | The method and apparatus of video frequency object tracking |
| CN107609497A (en) * | 2017-08-31 | 2018-01-19 | 武汉世纪金桥安全技术有限公司 | The real-time video face identification method and system of view-based access control model tracking technique |
| CN107644204A (en) * | 2017-09-12 | 2018-01-30 | 南京凌深信息科技有限公司 | A kind of human bioequivalence and tracking for safety-protection system |
| CN107679613A (en) * | 2017-09-30 | 2018-02-09 | 同观科技（深圳）有限公司 | A kind of statistical method of personal information, device, terminal device and storage medium |
| CN108171207A (en) * | 2018-01-17 | 2018-06-15 | 百度在线网络技术（北京）有限公司 | Face identification method and device based on video sequence |
| CN109815858A (en) * | 2019-01-10 | 2019-05-28 | 中国科学院软件研究所 | A kind of target user Gait Recognition system and method in surroundings |
| CN109919977A (en) * | 2019-02-26 | 2019-06-21 | 鹍骐科技（北京）股份有限公司 | A kind of video motion personage tracking and personal identification method based on temporal characteristics |
| CN110232323A (en) * | 2019-05-13 | 2019-09-13 | 特斯联（北京）科技有限公司 | A kind of parallel method for quickly identifying of plurality of human faces for crowd and its device |
| CN110941978A (en) * | 2019-05-23 | 2020-03-31 | 罗普特科技集团股份有限公司 | Face clustering method and device for unidentified personnel and storage medium |
| CN110503059A (en) * | 2019-08-27 | 2019-11-26 | 国网电子商务有限公司 | A kind of face identification method and system |
| CN111310731A (en) * | 2019-11-15 | 2020-06-19 | 腾讯科技（深圳）有限公司 | Video recommendation method, device and equipment based on artificial intelligence and storage medium |
| CN111079670A (en) * | 2019-12-20 | 2020-04-28 | 北京百度网讯科技有限公司 | Face recognition method, face recognition device, face recognition terminal and face recognition medium |
| CN111401171A (en) * | 2020-03-06 | 2020-07-10 | 咪咕文化科技有限公司 | Face image recognition method and device, electronic equipment and storage medium |
| CN111444817A (en) * | 2020-03-24 | 2020-07-24 | 咪咕文化科技有限公司 | Person image identification method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111147513B (en) | Transverse moving attack path determination method in honey net based on attack behavior analysis | |
| CN110334569B (en) | Passenger flow volume in-out identification method, device, equipment and storage medium | |
| CN110610127B (en) | Face recognition method and device, storage medium and electronic equipment | |
| CN111738120B (en) | Character recognition method, character recognition device, electronic equipment and storage medium | |
| CN111161206A (en) | Image capturing method, monitoring camera and monitoring system | |
| CN112637568B (en) | Distributed security monitoring method and system based on multi-node edge computing equipment | |
| CN110599129A (en) | Campus attendance checking method, device, identification terminal and system based on image tracking | |
| US11164327B2 (en) | Estimation of human orientation in images using depth information from a depth camera | |
| CN112784835A (en) | Method and device for identifying authenticity of circular seal, electronic equipment and storage medium | |
| CN111091106A (en) | Image clustering method and device, storage medium and electronic device | |
| CN113255549A (en) | Intelligent recognition method and system for pennisseum hunting behavior state | |
| CN112200084A (en) | Face recognition method and device for video stream, electronic equipment and storage medium | |
| CN116597421A (en) | Parking space monitoring method, device and equipment based on image recognition | |
| CN115375886A (en) | Data acquisition method and system based on cloud computing service | |
| CN114842393A (en) | Statistical method and device for pedestrian flow | |
| CN115457467A (en) | Building quality hidden danger positioning method and system based on data mining | |
| CN113554685A (en) | Method and device for detecting moving target of remote sensing satellite, electronic equipment and storage medium | |
| CN114332983A (en) | Face image definition detection method, face image definition detection device, electronic equipment and medium | |
| CN110020624B (en) | Image recognition method, terminal device and storage medium | |
| CN113361455A (en) | Training method of face counterfeit identification model, related device and computer program product | |
| CN113111847A (en) | Automatic monitoring method, device and system for process circulation | |
| CN113592427A (en) | Method and apparatus for counting man-hours and computer readable storage medium | |
| CN115273215A (en) | Job recognition system and job recognition method | |
| CN112819859A (en) | Multi-target tracking method and device applied to intelligent security | |
| CN112800816A (en) | Video motion recognition detection method based on multiple models |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 100083 Room 806, 8 / F, building 1, courtyard a 11, Anxiang Beili, Chaoyang District, Beijing Applicant after: HUAHANG HI-TECH (BEIJING) TECHNOLOGY Co.,Ltd. Address before: 100083 C-1103 room 18 Zhongguancun East Road, Haidian District, Beijing. Applicant before: HUAHANG HI-TECH (BEIJING) TECHNOLOGY Co.,Ltd. |