CN104794464B - A kind of biopsy method based on relative priority - Google Patents
A kind of biopsy method based on relative priority Download PDFInfo
- Publication number
- CN104794464B CN104794464B CN201510243778.XA CN201510243778A CN104794464B CN 104794464 B CN104794464 B CN 104794464B CN 201510243778 A CN201510243778 A CN 201510243778A CN 104794464 B CN104794464 B CN 104794464B
- Authority
- CN
- China
- Prior art keywords
- face
- relative priority
- key point
- change value
- judged
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
-
- 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
- G06V40/174—Facial expression recognition
- G06V40/176—Dynamic expression
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of biopsy methods based on relative priority, comprising the following steps: 1) detects the face location in input video on every frame image；2) face key point is obtained；3) region of eyes or mouth is obtained according to acquired key point；4) judge whether the rule of the attribute change value in the obtained region of step 3) in continuous multiple frames image meets the changing rule of real human face, if so, being judged as real human face, if it is not, being then judged as false face.Compared with prior art, the present invention has many advantages, such as that detection accuracy is high, speed is fast.
Description
Technical field
The present invention relates to a kind of human face detection tech, more particularly, to a kind of biopsy method based on relative priority.
Background technique
Recognition of face has succeeded as a kind of identity identifying technology in fields such as public security protection, attendance gate inhibitions.But
Conventional face's identification technology does not consider the true and false of target face, therefore is easy the attack by false face.If false face
Success attack, it is possible to heavy losses be caused to user, therefore reliable and efficient face In vivo detection technology is tested as face
The important component of card system.
Conventional face's identification technology is come commonly using the methods of Fourier analysis, blink detection and three dimensional depth estimation
Judge living body.But these method difficulties meet the requirement of finance and public safety-security area.Main there are two factors: 1) these algorithms
Hardly possible meets rate of false alarm less than one thousandth in performance and percent of pass will be more than 95% index；2) difficult to resist specific false face
Attack, as blink detection difficulty resist based on video false face attack.
Tradition close one's eyes (or opening one's mouth) classifier performance be difficult to meet demand, such as have human eye smaller or have human eye without
Method is closed completely.
Summary of the invention
High, speed that it is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide a kind of detection accuracy
The fast biopsy method based on relative priority of degree.
The purpose of the present invention can be achieved through the following technical solutions:
A kind of biopsy method based on relative priority, comprising the following steps:
1) face location in input video on every frame image is detected；
2) face key point is obtained；
3) region of eyes or mouth is obtained according to acquired key point；
4) judge whether the rule of the attribute change value in the obtained region of step 3) in continuous multiple frames image meets true people
The changing rule of face, if so, being judged as real human face, if it is not, being then judged as false face.
Distance change value of the attribute change value between upper lower eyelid or the distance change value between upper lower lip.
The step 4) specifically:
401) eyes or mouth region merging technique of present frame and preceding t frame are schemed at one, using returning based on deep learning
Return attribute change value in method output two field pictures；
402) step 401) is repeated until obtaining the attribute change value of every frame image；
403) by all properties changing value by frame time sequence form a vector, using SVM classifier to the vector into
Row classification；
404) judge whether classification results meet the changing rule of the real human face under set action, if so, being judged as
Real human face, if it is not, being then judged as false face.
In the step 403), before being classified using SVM classifier to the vector, the length of each vector is set
It is fixed.
The set action includes closing one's eyes or opening one's mouth.
The input video is one section of 3~5 seconds face video.
The step 1) uses AdaBoost detection of classifier face location.
In the step 2), the detailed process of face key point is obtained are as follows:
201) by HoG in conjunction with SVM in the way of carry out first round critical point detection, each key point has the selection of K kind；
202) global shape information is utilized, using N-Best mode, it is crucial that composition face shape is obtained in K^N kind is possible
Point optimal solution, N are key point number, carry out lopping processing using branch-and-bound mode, obtain final form families；
203) confidence level for calculating the every kind of combination obtained in step 202) chooses a high combination of confidence level.
The confidence level is made of two parts:
A) confidence level obtained in the way of HoG in step 201)；
B) positional relationship between different key points.
Compared with prior art, the invention has the following advantages that
1) present invention carries out living body faces detection using attribute change information, compared with prior art, has in performance obvious
It is promoted, meets rate of false alarm less than one thousandth and percent of pass will be more than 95% index；
2) algorithm used in the method for the present invention has fireballing advantage, improves the detection rates of face, and processing 3~
5 seconds videos only need 0.5 second time；
3) the small-sized client (such as smart phone) of the present invention can achieve live effect；
4) present invention carries out living body faces detection by attribute change information, can satisfy as eyes are smaller or eyes can not
The demand of occasions such as it is closed completely.
Detailed description of the invention
Fig. 1 is flow diagram of the invention.
Specific embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.The present embodiment is with technical solution of the present invention
Premised on implemented, the detailed implementation method and specific operation process are given, but protection scope of the present invention is not limited to
Following embodiments.
The present embodiment provides a kind of biopsy methods based on relative priority, by the way of human-computer interaction, allow inspection
Object is surveyed the specific actions such as to close one's eyes, open one's mouth, by judging whether test object completes these work, thus judge whether be
Real human face.As shown in Figure 1, this method the following steps are included:
Step S1, using the face location on every frame image in AdaBoost detection of classifier input video, input video
For one section of 3~5 seconds face video.
Step S2 obtains face key point, detailed process are as follows:
201) by HoG in conjunction with SVM in the way of carry out first round critical point detection, each key point has the selection of K kind；
202) global shape information is utilized, using N-Best mode, it is crucial that composition face shape is obtained in K^N kind is possible
Point optimal solution, N are key point number, carry out lopping processing using branch-and-bound mode, obtain final form families；
203) confidence level for calculating the every kind of combination obtained in step 202) chooses a high combination of confidence level.This is wanted
It asks and meets Gaussian Profile, which obtains according to statistics in advance.Confidence level is made of two parts: a) being utilized in step 201)
The confidence level that HoG mode obtains；B) positional relationship between different key points.
Step S3 obtains the region of eyes or mouth according to acquired key point.
It is true to judge whether the rule of the attribute change value in the obtained region of step 3) in continuous multiple frames image meets by step S4
The changing rule of real face, if so, it is judged as real human face, if it is not, then it is judged as false face, specifically:
401) eyes or mouth region merging technique of present frame and preceding t frame are schemed at one, using returning based on deep learning
Return attribute change value in method output two field pictures；
402) step 401) is repeated until obtaining the attribute change value of every frame image, the attribute change value is between upper lower eyelid
Distance change value or upper lower lip between distance change value；
403) by all properties changing value by frame time sequence form a vector, using SVM classifier to the vector into
Row classification；
SVM can only handle the feature vector of specific length, and due to the speed issue of human action, feature vector length is different
It causes.The vector of regular length is obtained using three kinds of modes: a) removing the frame of foremost；B) rearmost frame is removed；C) according to
Certain frequency sampling.
404) judge whether classification results meet the changing rule of the real human face under set action (such as close one's eyes or open one's mouth),
If so, being judged as real human face, if it is not, being then judged as false face.If it is eye closing, distance change value will appear by becoming greatly
Mode that is small, then changing from small to big, this is that false face attack can not simulate substantially.
Claims (7)
1. a kind of biopsy method based on relative priority, which comprises the following steps:
1) face location in input video on every frame image is detected；
2) face key point is obtained, the detailed process of face key point is obtained are as follows:
201) by HoG in conjunction with SVM in the way of carry out first round critical point detection, each key point has the selection of K kind；
202) global shape information is utilized, using N-Best mode, in KNIt is optimal that composition face shape key point is obtained during kind is possible
Solution, N are key point number, carry out lopping processing using branch-and-bound mode, obtain final form families；
203) confidence level for calculating the every kind of combination obtained in step 202) chooses a high combination of confidence level；
3) region of eyes or mouth is obtained according to acquired key point；
4) judge whether the rule of the attribute change value in the obtained region of step 3) in continuous multiple frames image meets real human face
Changing rule, if so, being judged as real human face, if it is not, being then judged as false face, wherein the attribute change value is upper
The distance change value between distance change value or upper lower lip between lower eyelid.
2. the biopsy method according to claim 1 based on relative priority, which is characterized in that the step 4) is specific
Are as follows:
401) eyes or mouth region merging technique of present frame and preceding t frame are schemed at one, using the recurrence side based on deep learning
Method exports attribute change value in two field pictures；
402) step 401) is repeated until obtaining the attribute change value of every frame image；
403) all properties changing value is formed into a vector by frame time sequence, the vector is divided using SVM classifier
Class；
404) judge whether classification results meet the changing rule of the real human face under set action, if so, being judged as true
Face, if it is not, being then judged as false face.
3. the biopsy method according to claim 2 based on relative priority, which is characterized in that the step 403)
In, before being classified using SVM classifier to the vector, the length of each vector is set.
4. the biopsy method according to claim 2 based on relative priority, which is characterized in that the set action packet
It includes eye closing or opens one's mouth.
5. the biopsy method according to claim 1 based on relative priority, which is characterized in that the input video is
One section of 3~5 seconds face video.
6. the biopsy method according to claim 1 based on relative priority, which is characterized in that the step 1) uses
AdaBoost detection of classifier face location.
7. the biopsy method according to claim 1 based on relative priority, which is characterized in that the confidence level is by two
Part is constituted:
A) confidence level obtained in the way of HoG in step 201)；
B) positional relationship between different key points.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510243778.XA CN104794464B (en) | 2015-05-13 | 2015-05-13 | A kind of biopsy method based on relative priority |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510243778.XA CN104794464B (en) | 2015-05-13 | 2015-05-13 | A kind of biopsy method based on relative priority |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104794464A CN104794464A (en) | 2015-07-22 |
| CN104794464B true CN104794464B (en) | 2019-06-07 |
Family
ID=53559251
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510243778.XA Active CN104794464B (en) | 2015-05-13 | 2015-05-13 | A kind of biopsy method based on relative priority |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104794464B (en) |
Families Citing this family (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105046227B (en) * | 2015-07-24 | 2018-07-31 | 上海依图网络科技有限公司 | A kind of key frame acquisition methods for portrait video system |
| CN106557723B (en) * | 2015-09-25 | 2020-01-24 | 北京市商汤科技开发有限公司 | Face identity authentication system with interactive living body detection and method thereof |
| CN106557726B (en) * | 2015-09-25 | 2020-06-09 | 北京市商汤科技开发有限公司 | Face identity authentication system with silent type living body detection and method thereof |
| CN105139503A (en) * | 2015-10-12 | 2015-12-09 | 北京航空航天大学 | Lip moving mouth shape recognition access control system and recognition method |
| CN105243378B (en) * | 2015-11-13 | 2019-03-01 | 清华大学 | Living body faces detection method and device based on eye information |
| CN107625527B (en) * | 2016-07-19 | 2021-04-20 | 杭州海康威视数字技术股份有限公司 | Lie detection method and device |
| CN107358152B (en) * | 2017-06-02 | 2020-09-08 | 广州视源电子科技股份有限公司 | Living body identification method and system |
| CN107330914B (en) * | 2017-06-02 | 2021-02-02 | 广州视源电子科技股份有限公司 | Human face part motion detection method and device and living body identification method and system |
| CN107330370B (en) * | 2017-06-02 | 2020-06-19 | 广州视源电子科技股份有限公司 | Forehead wrinkle action detection method and device and living body identification method and system |
| CN109389002A (en) * | 2017-08-02 | 2019-02-26 | 阿里巴巴集团控股有限公司 | Biopsy method and device |
| CN108171158B (en) * | 2017-12-27 | 2022-05-17 | 北京迈格威科技有限公司 | Living body detection method, living body detection device, electronic apparatus, and storage medium |
| WO2019127262A1 (en) * | 2017-12-28 | 2019-07-04 | 深圳前海达闼云端智能科技有限公司 | Cloud end-based human face in vivo detection method, electronic device and program product |
| CN110032915A (en) * | 2018-01-12 | 2019-07-19 | 杭州海康威视数字技术股份有限公司 | A kind of human face in-vivo detection method, device and electronic equipment |
| CN108487820B (en) * | 2018-03-07 | 2019-12-27 | 常州信息职业技术学院 | Electronic detection closing device |
| CN109271929B (en) * | 2018-09-14 | 2020-08-04 | 北京字节跳动网络技术有限公司 | Detection method and device |
| CN109492551B (en) | 2018-10-25 | 2023-03-24 | 腾讯科技（深圳）有限公司 | Living body detection method and device and related system applying living body detection method |
| CN109658533A (en) * | 2018-11-23 | 2019-04-19 | 深圳市沃特沃德股份有限公司 | Method of registering, system and the intelligent terminal of video conference |
| CN109522863B (en) * | 2018-11-28 | 2020-11-27 | 北京达佳互联信息技术有限公司 | Ear key point detection method and device and storage medium |
| CN109697416B (en) * | 2018-12-14 | 2022-11-18 | 腾讯科技（深圳）有限公司 | Video data processing method and related device |
| CN109858402B (en) * | 2019-01-16 | 2021-08-31 | 腾讯科技（深圳）有限公司 | Image detection method, device, terminal and storage medium |
| CN110059624B (en) * | 2019-04-18 | 2021-10-08 | 北京字节跳动网络技术有限公司 | Method and apparatus for detecting living body |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103440479A (en) * | 2013-08-29 | 2013-12-11 | 湖北微模式科技发展有限公司 | Method and system for detecting living body human face |
| CN104361573A (en) * | 2014-09-26 | 2015-02-18 | 北京航空航天大学 | Color information and global information fused SIFT (scale invariant feature transform) feature matching algorithm |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102360421B (en) * | 2011-10-19 | 2014-05-28 | 苏州大学 | Face identification method and system based on video streaming |
| US9025830B2 (en) * | 2012-01-20 | 2015-05-05 | Cyberlink Corp. | Liveness detection system based on face behavior |
| CN102646190B (en) * | 2012-03-19 | 2018-05-08 | 深圳市腾讯计算机系统有限公司 | A kind of authentication method based on biological characteristic, apparatus and system |
-
2015
- 2015-05-13 CN CN201510243778.XA patent/CN104794464B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103440479A (en) * | 2013-08-29 | 2013-12-11 | 湖北微模式科技发展有限公司 | Method and system for detecting living body human face |
| CN104361573A (en) * | 2014-09-26 | 2015-02-18 | 北京航空航天大学 | Color information and global information fused SIFT (scale invariant feature transform) feature matching algorithm |
Non-Patent Citations (3)
| Title |
|---|
| Face recognition with liveness detection using eye and mouth movement;Avinash Kumar Singh et al.;《IEEE Xplore Digital Library》;20140828;全文 * |
| The N-Best Algorithm：An Efficient Procedure for Finding Top N Sentence Hypotheses;Yen-Lu Chow et al.;《HLT"89 Proceedings of the workshop on Speech and Natural Language》;19891015;全文 * |
| 人脸识别中的活体检测技术研究;孙霖;《万方数据库》;20141103;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104794464A (en) | 2015-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104794464B (en) | A kind of biopsy method based on relative priority | |
| CN106557723B (en) | Face identity authentication system with interactive living body detection and method thereof | |
| CN100592322C (en) | An automatic computer authentication method for photographic faces and living faces | |
| CN106557726B (en) | Face identity authentication system with silent type living body detection and method thereof | |
| Hofmann et al. | The TUM Gait from Audio, Image and Depth (GAID) database: Multimodal recognition of subjects and traits | |
| Pan et al. | Eyeblink-based anti-spoofing in face recognition from a generic webcamera | |
| CN108182409B (en) | Living body detection method, living body detection device, living body detection equipment and storage medium | |
| CN108140123A (en) | Face living body detection method, electronic device and computer program product | |
| Kawulok et al. | Precise multi-level face detector for advanced analysis of facial images | |
| US8965068B2 (en) | Apparatus and method for discriminating disguised face | |
| Rao et al. | Sign Language Recognition System Simulated for Video Captured with Smart Phone Front Camera. | |
| CN103778409A (en) | Human face identification method based on human face characteristic data mining and device | |
| TWI603293B (en) | Method and apparatus for detecting blink | |
| WO2013075295A1 (en) | Clothing identification method and system for low-resolution video | |
| CN104102903A (en) | Secondary face recognition method based on SRC | |
| Putro et al. | Adult image classifiers based on face detection using Viola-Jones method | |
| CN107977622A (en) | Eyes detection method based on pupil feature | |
| US10068127B2 (en) | Automatic detection of face and thereby localize the eye region for iris recognition | |
| Gul et al. | A machine learning approach to detect occluded faces in unconstrained crowd scene | |
| Nguyen et al. | Towards recognizing facial expressions at deeper level: Discriminating genuine and fake smiles from a sequence of images | |
| Nanthini et al. | An Efficient Velocity Estimation Approach for Face Liveness Detection using Sparse Optical Flow Technique | |
| KR101342018B1 (en) | Real-time Object Recognition and Tracking Method Using Representative Feature, and Apparatus Thereof | |
| Hazar et al. | Real time face detection based on motion and skin color information | |
| Mariappan et al. | A labVIEW design for frontal and non-frontal human face detection system in complex background | |
| KR102674496B1 (en) | Method and Apparatus for Detecting Fake Faces Using Remote Photoplethysmography Signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190919 Address after: Room 1901E, 488 Yaohua Road, Pudong New Area, Shanghai 201125 Patentee after: Shanghai Yitu Information Technology Co., Ltd. Address before: 200240, room 1261, South Building, No. 4915 TV Park, North Pine Road, Songjiang District, Shanghai Patentee before: Shanghai is according to figure network technology company limited |