CN110287933A - A kind of dynamic human face recognition system and recognition methods based on stereo video streaming - Google Patents
A kind of dynamic human face recognition system and recognition methods based on stereo video streaming Download PDFInfo
- Publication number
- CN110287933A CN110287933A CN201910588589.4A CN201910588589A CN110287933A CN 110287933 A CN110287933 A CN 110287933A CN 201910588589 A CN201910588589 A CN 201910588589A CN 110287933 A CN110287933 A CN 110287933A
- Authority
- CN
- China
- Prior art keywords
- face
- module
- recognition
- tracking
- video streaming
- 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- Withdrawn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
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
- G06V40/161—Detection; Localisation; Normalisation
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of dynamic human face recognition system based on stereo video streaming and recognition methods, including video flowing, face detection module, face tracking module, face cluster module, face recognition module, the output end of the video flowing is electrically connected with the input terminal of face detection module and face tracking module respectively, and the output end of face detection module and the input terminal of face tracking module are electrically connected, the output end of the face tracking module and the input terminal of face cluster module are electrically connected.In the present invention, using multi-tag classification and multitask recognition methods, the key point information and face character information of face are identified while detecting face, the relevant information of face character is identified at the same time of face detection, the calculating time of Attribute Recognition is saved, proposes the clustering method based on tracking result, rejects the face picture of non-present id in tracking result, reduce manually to delete and select work, is conducive to the collection work of later period human face data.
Description
Technical field
The present invention relates to can technical field of face recognition more particularly to a kind of dynamic human face identification based on stereo video streaming
System.
Background technique
The research of face identification system starts from the 1960s, with computer technology and optical imagery skill after the eighties
The development of art is improved, and actually enters the primary application stage then 90 year later period, and with the U.S., Germany and Japan
Based on technology is realized；The successful key of face identification system is whether possess the core algorithm at tip, and has recognition result
There are practical discrimination and recognition speed；" face identification system " is integrated with artificial intelligence, machine recognition, machine learning, mould
A variety of professional techniques such as type theory, expert system, video image processing, while in conjunction with the theoretical of median processing and need to realize,
It is the more recent application of living things feature recognition, the realization of core technology presents conversion of the weak artificial intelligence to strong artificial intelligence.
Existing video flowing face dynamic recognition system mainly includes Face datection, and face tracking, picture quality determines, special
Sign is extracted, and aspect ratio peer modules, Face datection algorithm only extracts the face location information in picture at present, for face character
Information (it is whether fuzzy, whether band sunglasses, whether block) have no identification in the detection process, but the individual quality of use is sentenced
Algorithm is determined to identify face information, will cause compute repeatedly in this way, increases algorithm time-consuming, when blocking, replacing occur in face, with
Track algorithm can fail, and cause the same id in the result of tracking different people occur, before carrying out feature extraction, sentenced using quality
Determine algorithm and find out the best picture of picture quality in face tracking result to carry out feature extraction, but before the picture does not merge
The information of frame afterwards can have an impact to the accuracy rate of identification.
Summary of the invention
The purpose of the present invention is to solve disadvantages existing in the prior art, and the one kind proposed is based on stereo video streaming
Dynamic human face recognition system.
To achieve the goals above, present invention employs following technical solutions: a kind of dynamic people based on stereo video streaming
Face identifying system, including video flowing, face detection module, face tracking module, face cluster module, face recognition module；
The output end of the video flowing is electrically connected with the input terminal of face detection module and face tracking module respectively, and
The output end of face detection module and the input terminal of face tracking module are electrically connected, the output end of the face tracking module with
The input terminal of face cluster module is electrically connected, the output end of the face cluster module and the input terminal electricity of face recognition module
Property connection.
It is as above-mentioned technical proposal to further describe:
Face recognition module is by face location information module, face key point information module and face character information module group
At.
It is as above-mentioned technical proposal to further describe:
Face detection module is made of convolution residual error network, split tunnel convolution, for extracting characteristics of image, the network
Branched portion mainly has the output of three parts as output: the location information of face, the key point information of face and face its
His attribute information.
A kind of dynamic human face recognition methods based on stereo video streaming, comprising the following steps:
S1: identifying the image transmitting in video flowing to face detection module, if there is face then to face into
Row alignment and identification, if there is no then restarting to detect；
S2: after completing recognition of face, known face, then be tracked the face in video flowing, if it does not exist if it exists
Known face then intercepts facial image and is saved and restart to detect.
S3: face is birdsed of the same feather flock together: for the image of face tracking module acquisition, extracting people in tracking result with deep learning network
The feature of face image judges the class number in current tracking result using clustering algorithm, if class number is greater than 1, deletes
Except the less classification of picture number, guarantee only one classification of current results, and retention class center compares for identification.
It is as above-mentioned technical proposal to further describe:
The class center has merged the human face image information of all tracking results, uses this feature as recognition of face
The accuracy rate of identification can be improved in feature.
It is as above-mentioned technical proposal to further describe:
The clustering algorithm are as follows: judge the class number in current tracking result, if class number is greater than 1, delete
The less classification of picture number guarantees only one classification of current results, and retention class center compares for identification.
Beneficial effect
The present invention provides a kind of dynamic human face recognition systems based on stereo video streaming.Have it is following the utility model has the advantages that
(1) using multi-tag classification and multitask recognition methods, the key of face is identified while detecting face
Point information and face character information, identify the relevant information of face character at the same time of face detection, save attribute knowledge
Other calculating time；
(2) it proposes the clustering method based on tracking result, rejects the face picture of non-present id in tracking result, subtract
Lack manually to delete and selected work, has been conducive to the collection work of later period human face data；
(3) face identification method based on cluster centre is proposed, this method is using the class center of tracking result as people
Face identification feature, so that the feature of before and after frames facial image has been merged, the fusion of before and after frames human face image information, to promote people
The accuracy rate of face identification.
Detailed description of the invention
Fig. 1 is a kind of Face datection and attribute of the dynamic human face recognition system based on stereo video streaming proposed by the present invention
Recognition methods flow diagram；
Fig. 2 is the clustering method flow diagram based on tracking result in the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.
Referring to Fig.1-2, a kind of dynamic human face recognition system based on stereo video streaming, including video flowing, Face datection mould
Block, face tracking module, face cluster module, face recognition module；
The output end of the video flowing is electrically connected with the input terminal of face detection module and face tracking module respectively, and
The output end of face detection module and the input terminal of face tracking module are electrically connected, the output end of the face tracking module with
The input terminal of face cluster module is electrically connected, the output end of the face cluster module and the input terminal electricity of face recognition module
Property connection.
Face recognition module is by face location information module, face key point information module and face character information module group
At, face detection module is made of convolution residual error network, split tunnel convolution, for extracting characteristics of image, the branch of the network
Mainly there is the output of three parts: other categories of the location information of face, the key point information of face and face in part as output
Property information.
A kind of dynamic human face recognition methods based on stereo video streaming, comprising the following steps:
S1: identifying the image transmitting in video flowing to face detection module, if there is face then to face into
Row alignment and identification, if there is no then restarting to detect；
S2: after completing recognition of face, known face, then be tracked the face in video flowing, if it does not exist if it exists
Known face then intercepts facial image and is saved and restart to detect.
S3: face is birdsed of the same feather flock together: for the image of face tracking module acquisition, extracting people in tracking result with deep learning network
The feature of face image judges the class number in current tracking result using clustering algorithm, if class number is greater than 1, deletes
Except the less classification of picture number, guarantee only one classification of current results, and retention class center compares for identification.
Class center has merged the human face image information of all tracking results, uses this feature as the feature of recognition of face
The accuracy rate of identification, clustering algorithm can be improved are as follows: judge the class number in current tracking result, if class number is greater than 1,
The less classification of picture number is then deleted, guarantees only one classification of current results, and retention class center compares for identification.
The present invention identifies the pass of face using multi-tag classification and multitask recognition methods while detecting face
Key point information and face character information, identify the relevant information of face character at the same time of face detection, save attribute
The calculating time of identification proposes the clustering method based on tracking result, rejects the face figure of non-present id in tracking result
Piece reduces manually to delete and selects work, is conducive to the collection work of later period human face data, proposes the recognition of face based on cluster centre
Method, this method is using the class center of tracking result as face recognition features, to merge before and after frames facial image
Feature, the fusion of before and after frames human face image information, to promote the accuracy rate of recognition of face.
In the description of this specification, the description of reference term " one embodiment ", " example ", " specific example " etc. means
Specific features described in conjunction with this embodiment or example, structure, material live feature and are contained at least one implementation of the invention
In example or example.In the present specification, schematic expression of the above terms may not refer to the same embodiment or example.
Moreover, particular features, structures, materials, or characteristics described can be in any one or more of the embodiments or examples to close
Suitable mode combines.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (6)
1. a kind of dynamic human face recognition system based on stereo video streaming, which is characterized in that including video flowing, Face datection mould
Block, face tracking module, face cluster module, face recognition module；
The output end of the video flowing is electrically connected with the input terminal of face detection module and face tracking module respectively, and face
The output end of detection module and the input terminal of face tracking module are electrically connected, the output end and face of the face tracking module
The input terminal of cluster module is electrically connected, and the output end of the face cluster module and the input terminal of face recognition module electrically connect
It connects.
2. a kind of dynamic human face recognition system based on stereo video streaming according to claim 1, which is characterized in that described
Face recognition module is made of face location information module, face key point information module and face character information module.
3. a kind of dynamic human face recognition system based on stereo video streaming according to claim 1, which is characterized in that described
Face detection module is made of convolution residual error network, split tunnel convolution, for extracting characteristics of image, the branched portion of the network
As output, mainly there is the output of three parts: other attributes letter of the location information of face, the key point information of face and face
Breath.
4. a kind of dynamic human face recognition methods based on stereo video streaming, which comprises the following steps:
S1: identifying the image transmitting in video flowing to face detection module, then carries out pair to face if there is face
Neat and identification, if there is no then restarting to detect；
S2: after completing recognition of face, known face, then be tracked the face in video flowing if it exists, known if it does not exist
Face then intercepts facial image and is saved and restart to detect.
S3: face is birdsed of the same feather flock together: for the image of face tracking module acquisition, extracting face figure in tracking result with deep learning network
The feature of picture judges the class number in current tracking result using clustering algorithm, if class number is greater than 1, deletes figure
As small numbers of classification, guarantee only one classification of current results, and retention class center compares for identification.
5. a kind of dynamic human face recognition methods based on stereo video streaming according to claim 4, which is characterized in that described
Class center has merged the human face image information of all tracking results, uses this feature that knowledge can be improved as the feature of recognition of face
Other accuracy rate.
6. a kind of dynamic human face recognition methods based on stereo video streaming according to claim 4, which is characterized in that described
Clustering algorithm are as follows: judge the class number in current tracking result, if class number is greater than 1, it is less to delete picture number
Classification, guarantee only one classification of current results, and retention class center compares for identification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910588589.4A CN110287933A (en) | 2019-07-02 | 2019-07-02 | A kind of dynamic human face recognition system and recognition methods based on stereo video streaming |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910588589.4A CN110287933A (en) | 2019-07-02 | 2019-07-02 | A kind of dynamic human face recognition system and recognition methods based on stereo video streaming |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110287933A true CN110287933A (en) | 2019-09-27 |
Family
ID=68021649
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910588589.4A Withdrawn CN110287933A (en) | 2019-07-02 | 2019-07-02 | A kind of dynamic human face recognition system and recognition methods based on stereo video streaming |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110287933A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111898467A (en) * | 2020-07-08 | 2020-11-06 | 浙江大华技术股份有限公司 | Attribute identification method and device, storage medium and electronic device |
| CN114120506A (en) * | 2021-09-30 | 2022-03-01 | 国网浙江省电力有限公司 | Infrastructure field personnel management and control system and method based on 5G network architecture |
-
2019
- 2019-07-02 CN CN201910588589.4A patent/CN110287933A/en not_active Withdrawn
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111898467A (en) * | 2020-07-08 | 2020-11-06 | 浙江大华技术股份有限公司 | Attribute identification method and device, storage medium and electronic device |
| CN111898467B (en) * | 2020-07-08 | 2023-02-28 | 浙江大华技术股份有限公司 | Attribute identification method and device, storage medium and electronic device |
| CN114120506A (en) * | 2021-09-30 | 2022-03-01 | 国网浙江省电力有限公司 | Infrastructure field personnel management and control system and method based on 5G network architecture |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106295568B (en) | The mankind's nature emotion identification method combined based on expression and behavior bimodal | |
| CN106778664B (en) | Iris image iris area segmentation method and device | |
| CN108829900B (en) | Face image retrieval method and device based on deep learning and terminal | |
| CN111339847B (en) | Face emotion recognition method based on graph convolution neural network | |
| CN109697416A (en) | A kind of video data handling procedure and relevant apparatus | |
| CN111563452B (en) | Multi-human-body gesture detection and state discrimination method based on instance segmentation | |
| CN105160318A (en) | Facial expression based lie detection method and system | |
| CN106845373A (en) | Towards pedestrian's attribute forecast method of monitor video | |
| CN103618918A (en) | Method and device for controlling display of smart television | |
| CN109522853A (en) | Face datection and searching method towards monitor video | |
| CN109670405A (en) | A kind of complex background pedestrian detection method based on deep learning | |
| CN109753904A (en) | A kind of face identification method and system | |
| CN113627402B (en) | Image identification method and related device | |
| CN110008793A (en) | Face identification method, device and equipment | |
| CN110287933A (en) | A kind of dynamic human face recognition system and recognition methods based on stereo video streaming | |
| CN113920568A (en) | Face and human body posture emotion recognition method based on video image | |
| CN111862413A (en) | Method and system for realizing epidemic situation resistant non-contact multidimensional identity rapid identification | |
| CN110176025A (en) | A kind of proctor's tracking based on posture | |
| CN113591692A (en) | Multi-view identity recognition method | |
| CN110008876A (en) | A kind of face verification method based on data enhancing and Fusion Features | |
| CN117333908A (en) | Cross-modal pedestrian re-recognition method based on attitude feature alignment | |
| CN103366163A (en) | Human face detection system and method based on incremental learning | |
| CN111694980A (en) | Robust family child learning state visual supervision method and device | |
| CN111523461A (en) | Expression recognition system and method based on enhanced CNN and cross-layer LSTM | |
| CN110135362A (en) | A kind of fast face recognition method based under infrared camera |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190927 |