CN109063533A - A kind of dynamic face Fast Recognition Algorithm - Google Patents
A kind of dynamic face Fast Recognition Algorithm Download PDFInfo
- Publication number
- CN109063533A CN109063533A CN201810406470.6A CN201810406470A CN109063533A CN 109063533 A CN109063533 A CN 109063533A CN 201810406470 A CN201810406470 A CN 201810406470A CN 109063533 A CN109063533 A CN 109063533A
- Authority
- CN
- China
- Prior art keywords
- face
- image
- recognition
- carried out
- lim
- 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
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
- G06V20/00—Scenes; Scene-specific elements
- G06V20/30—Scenes; Scene-specific elements in albums, collections or shared content, e.g. social network photos or video
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of dynamic face Fast Recognition Algorithms, obtain t moment face recognition result, it is extracted by score characteristic variations value SRi to image frame sequence and loading characteristic changing value TRi, it is compared with critical score value changing value vlim and critical load changing value wlim, as vlim > max (SRi) and wlim > min (TRi), face recognition result in the l duration is all made of the result of t moment recognition of face, otherwise, carry out the recognition of face of a new round, in practical applications, real-time face identification can be carried out to target person, and it is prompted in time in the higher comparison result of discovery similarity.
Description
Technical field
The present invention relates to a kind of recognition of face, specifically a kind of dynamic face Fast Recognition Algorithm.
Background technique
In social stability maintenance work, some known a suspects can be hidden in crowd, the police of public security system
It needs to scan for it in crowd and identify.In current identification, mainly by the way of manually interrogating and examining, borrow
Identity card reader is helped to complete the identification to target person, although this mode can give full play to a line police
Working experience abundant, still, since examination scope is wide, investigation officer quantity is big, will cause huge workload, needs
It is put into using a large amount of police strength.Moreover, checking by manpower, it is extremely easy to appear investigation careless omission.In addition, in police to people
In the troubleshooting procedure one by one of group, the vigilance of genuine suspicion people can be caused and absconded again, prevention of crimes is unfavorable for.
In recent years, with the development of technology, in existing technology, Video Supervision Technique is widely used in social safety dimension
In steady.But the effect of existing video monitoring is mainly reflected in and carries out subsequent analysis to video image, can not early warning in advance,
Its prevention and control ability is limited.Therefore, it during maintaining social stability and national security, needs to introduce first effective in turn
Technological means, ability of the General Promotion to all kinds of cases.
Summary of the invention
The purpose of the present invention is to provide a kind of dynamic face Fast Recognition Algorithms, to solve to mention in above-mentioned background technique
Out the problem of.
To achieve the above object, the invention provides the following technical scheme:
A kind of dynamic face Fast Recognition Algorithm, includes the following steps:
Step1: acquisition image;
Step2: image is pre-processed;
Step3: the face Gabor characteristic in image is obtained;
Step3: recognition of face is carried out based on Gabor characteristic, obtains recognition result RG;
Step4: dimension transformation is carried out to the image frame sequence (M, N, l) of acquisition, obtains two-dimensional matrix ((M+N), l)
Step5: carrying out PCA analysis, obtains score variation characteristic SRi and load variation characteristic Tri;
Step6: judgement v is carried outlim>max(SRi) and wlim>min(TRi), positive result is obtained, which is all made of identification
As a result RG;
Step7: updating two-dimensional matrix [(M+N), 1 (t1~tL+1)], PCA is iterated to calculate, into Step5;
Step8: judgement v is carried outlim>max(SRi) and wlim>min(TRi), obtain negative decision, into Step3, again into
Row recognition of face.
As the present invention further scheme: described to carry out that pretreated steps are as follows to image: the first step is to image
Binary conversion treatment is carried out, the face of people and profile each region of face are separated;Second step carries out gray processing processing.
Compared with prior art, the beneficial effects of the present invention are: the present invention obtains t moment face recognition result, by right
The score characteristic variations value SRi and loading characteristic changing value TRi of image frame sequence are extracted, with critical score value changing value
Vlim and critical load changing value wlim are compared, as vlim > max (SRi) and wlim > min (TRi), in the l duration
Face recognition result be all made of t moment recognition of face as a result, otherwise, the recognition of face of a new round is carried out, in practical application
In, real-time face identification can be carried out to target person, and prompt in time in the higher comparison result of discovery similarity.
Detailed description of the invention
Fig. 1 is the flow chart figure of dynamic face Fast Recognition Algorithm.
Fig. 2 is the facial image frame sequence figure acquired in dynamic face Fast Recognition Algorithm.
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.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts it is all its
His embodiment, shall fall within the protection scope of the present invention.
Please refer to Fig. 1~2, in the embodiment of the present invention, a kind of dynamic face Fast Recognition Algorithm includes the following steps:
Step1: acquisition image;
Step2: image is pre-processed;
Step3: the face Gabor characteristic in image is obtained;
Step3: recognition of face is carried out based on Gabor characteristic, obtains recognition result RG;
Step4: dimension transformation is carried out to the image frame sequence (M, N, l) of acquisition, obtains two-dimensional matrix ((M+N), l)
Step5: carrying out PCA analysis, obtains score variation characteristic SRi and load variation characteristic Tri;
Step6: judgement v is carried outlim>max(SRi) and wlim>min(TRi), positive result is obtained, which is all made of identification
As a result RG;
Step7: updating two-dimensional matrix [(M+N), 1 (t1~tL+1)], PCA is iterated to calculate, into Step5;
Step8: judgement v is carried outlim>max(SRi) and wlim>min(TRi), obtain negative decision, into Step3, again into
Row recognition of face.
Described to carry out that pretreated steps are as follows to image: the first step carries out binary conversion treatment to image, the face of people
It is separated with profile each region of face;Second step carries out gray processing processing.
By video acquisition to face video can regard image frame sequence as, the image of acquisition is located in advance first
Reason.The first step carries out binary conversion treatment to it, and the face of people and profile each region of face are separated, subsequent people is facilitated
The identification of face various pieces;Second step carries out gray processing processing, and grayscale image can greatly reduce complex background compared to cromogram
Influence, so that identification process is simplified.
Feature extraction is carried out to pretreated face picture, obtains the information that can uniquely correspond to a face.It is common
Feature include LBP, HOG, Gabor characteristic etc..Wherein, a kind of office with brightness invariance, description image of LBP feature
The feature of portion's unity and coherence in writing variation, this feature can be good at showing original image.HOG feature can describe face well
Part and Global Information, it is extremely sensitive to illumination and geographical location variation, if being used to the marginal information of picture engraving.
Gabor characteristic can arrive space, the selectivity in direction and frequency acquisition, the extraction especially suitable for multidate information.The present invention
Recognition of face is carried out using Gabor characteristic, obtains the face recognition result of single-frame images.
The basic principle of PCA analysis be one group is found from original data vector space being capable of utmostly characterize data side
The vector of difference, and a special vector matrix is constructed, so that initial data is mapped to low-dimensional feature sky from high-dimensional vector space
Between to realizing linear dimensionality reduction.
The facial image frame sequence of acquisition is as shown in Figure 2, wherein single-frame images is the matrix of M × N, and l is the time of image
Sequence.Dimension-reduction treatment is carried out to three-dimensional face sequence first, obtains two dimensional PCA matrix;Then two-dimensional matrix decompose and be obtained
Take pivot information;Score distribution map and loading characteristic curve are finally drawn according to pivot information, thus to the time of face sequence
It is analyzed with spatial signature information.It is specifically described the PCA Processing Algorithm basic principle of human face image sequence below.
For single width facial image, data structure is detailed in formula (1), if in human face image sequence including l frame face
Image then constitutes the three-dimensional matrice of (m × n × l) in Fig. 1.(m × n × l) three-dimensional matrice is subjected to dimension-reduction treatment, is turned
It is changed to ((m × n) × l) two-dimensional matrix TPCA:
PCA algorithm is by raw data matrix TPCAIt is disassembled by its covariance structure, is carried out with the linear combination of new variables
It indicates:
Wherein, X ∈ R(m×n)×l, P ∈ Rl×lScore vector matrix and load vector matrix are respectively indicated, for facial image
For sequence, the pivot and its projecting direction of human face image sequence are respectively indicated.Xi corresponds to i-th of score in score matrix X
Vector, score of each element representation different moments measurement data in principal component space.I-th in the corresponding load vector matrix P of Pi
Vector is loaded, indicates the mapping direction of i-th of pivot.
Score vector matrix and load vector matrix in PCA algorithm can be by the covariances of original human face image sequence matrix
Matrix is calculated:
In formula, R is covariance matrix,For TPCAThrough equalization treated result.Covariance matrix R is carried out special
Value indicative decomposition can solve load vector:
R=P Λ PT (4)
In formula, matrix Λ includes the non-negative characteristic value that amplitude is successively decreased, P=[p1,p2,…,pl] it is load matrix, p1、p2With
plRespectively first, second and l pivot load vector.And then it can be by X=TPCAP calculates each pivot score matrix.Finally
Shot chart and loading characteristic curve, shot chart and the load of facial image pivot are corresponded to load vector-drawn by score matrix
Characteristic curve respectively corresponds the spatial variations characteristic and time behavior of image change, can be described as:
In formula, SRiIndicate that image frame sequence corresponds to the spatial dimension of pivot, TRiFor time range.It can be respectively set critical
Score value changing value vlimWith critical load changing value wlim。
If the pivot shot chart and loading curve within the scope of corresponding room and time are continuous, i.e. vlim>max(SRi)
And wlim>min(TRi), then explanation in the image frame sequence collection process, image main feature there is no acute variation,
Then illustrate that target person and t moment are before same personage in image, at this point it is possible to directly know using the face before t moment
Other result.It is greater than threshold value when obtaining the intrinsic variation range with loading characteristic, then re-starts recognition of face.It in this way can pole
It is big to save the recognition of face time.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case where without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, nothing
By from the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by institute
Attached claim rather than above description limit, it is intended that will fall within the meaning and scope of the equivalent elements of the claims
All changes be included within the present invention.It should not treat any reference in the claims as limiting related right
It is required that.
In addition, it should be understood that although this specification is described in terms of embodiments, but not each embodiment is only wrapped
Containing an independent technical solution, this description of the specification is merely for the sake of clarity, and those skilled in the art should
It considers the specification as a whole, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art
The other embodiments being understood that.
Claims (2)
1. a kind of dynamic face Fast Recognition Algorithm, which comprises the steps of:
Step1: acquisition image;
Step2: image is pre-processed;
Step3: the face Gabor characteristic in image is obtained;
Step3: recognition of face is carried out based on Gabor characteristic, obtains recognition result RG;
Step4: dimension transformation is carried out to the image frame sequence (M, N, l) of acquisition, obtains two-dimensional matrix ((M+N), l)
Step5: carrying out PCA analysis, obtains score variation characteristic SRi and load variation characteristic Tri;
Step6: judgement v is carried outlim>max(SRi) and wlim>min(TRi), positive result is obtained, which is all made of recognition result
RG;
Step7: updating two-dimensional matrix [(M+N), 1 (t1 ~ tL+1)], PCA is iterated to calculate, into Step5;
Step8: judgement v is carried outlim>max(SRi) and wlim>min(TRi), negative decision is obtained, into Step3, re-starts people
Face identification.
2. dynamic face Fast Recognition Algorithm according to claim 1, which is characterized in that described to be located in advance to image
The step of reason is as follows: the first step carries out binary conversion treatment to image, and the face of people and profile each region of face are separated;
Second step carries out gray processing processing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810406470.6A CN109063533A (en) | 2018-04-30 | 2018-04-30 | A kind of dynamic face Fast Recognition Algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810406470.6A CN109063533A (en) | 2018-04-30 | 2018-04-30 | A kind of dynamic face Fast Recognition Algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109063533A true CN109063533A (en) | 2018-12-21 |
Family
ID=64819974
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810406470.6A Pending CN109063533A (en) | 2018-04-30 | 2018-04-30 | A kind of dynamic face Fast Recognition Algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109063533A (en) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101534305A (en) * | 2009-04-24 | 2009-09-16 | 中国科学院计算技术研究所 | Method and system for detecting network flow exception |
CN101710382A (en) * | 2009-12-07 | 2010-05-19 | 深圳大学 | Gabor human face recognizing method based on simplified intelligent single-particle optimizing algorithm |
CN101936971A (en) * | 2010-09-13 | 2011-01-05 | 天津大学 | Method for seeking biomarkers in penicillin fermentation process |
CN102306290A (en) * | 2011-10-14 | 2012-01-04 | 刘伟华 | Face tracking recognition technique based on video |
CN103310200A (en) * | 2013-06-25 | 2013-09-18 | 郑州吉瑞特电子科技有限公司 | Face recognition method |
CN104517104A (en) * | 2015-01-09 | 2015-04-15 | 苏州科达科技股份有限公司 | Face recognition method and face recognition system based on monitoring scene |
CN105975757A (en) * | 2016-04-28 | 2016-09-28 | 彩虹无线(北京)新技术有限公司 | Urgent speed reduction behavior recognition method based on vehicle driving data |
CN106500846A (en) * | 2016-09-22 | 2017-03-15 | 电子科技大学 | A kind of asymmetric correction method of infrared imaging system |
CN107203150A (en) * | 2017-05-22 | 2017-09-26 | 西安电子科技大学 | Asymmetric correction method based on infrared semi-matter simulating system |
-
2018
- 2018-04-30 CN CN201810406470.6A patent/CN109063533A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101534305A (en) * | 2009-04-24 | 2009-09-16 | 中国科学院计算技术研究所 | Method and system for detecting network flow exception |
CN101710382A (en) * | 2009-12-07 | 2010-05-19 | 深圳大学 | Gabor human face recognizing method based on simplified intelligent single-particle optimizing algorithm |
CN101936971A (en) * | 2010-09-13 | 2011-01-05 | 天津大学 | Method for seeking biomarkers in penicillin fermentation process |
CN102306290A (en) * | 2011-10-14 | 2012-01-04 | 刘伟华 | Face tracking recognition technique based on video |
CN103310200A (en) * | 2013-06-25 | 2013-09-18 | 郑州吉瑞特电子科技有限公司 | Face recognition method |
CN104517104A (en) * | 2015-01-09 | 2015-04-15 | 苏州科达科技股份有限公司 | Face recognition method and face recognition system based on monitoring scene |
CN105975757A (en) * | 2016-04-28 | 2016-09-28 | 彩虹无线(北京)新技术有限公司 | Urgent speed reduction behavior recognition method based on vehicle driving data |
CN106500846A (en) * | 2016-09-22 | 2017-03-15 | 电子科技大学 | A kind of asymmetric correction method of infrared imaging system |
CN107203150A (en) * | 2017-05-22 | 2017-09-26 | 西安电子科技大学 | Asymmetric correction method based on infrared semi-matter simulating system |
Non-Patent Citations (5)
Title |
---|
HAZEM M.EL-BAKRY: "New Fast Principal Component Analysis for Face Detection", 《RESEARCHGATE》 * |
方文东等: "海洋大数据序列时空变化主成分分析方法", 《热带海洋》 * |
杨国亮等: "基于帧间相似性约束鲁棒主成分分析模型的运动目标检测", 《计算机应用与软件》 * |
邵大培等: "基于PCA和图像匹配的飞机识别算法", 《中国体视学与图像分析》 * |
郭欣: "基于有监督Kohonen神经网络的步态识别", 《自动化学报》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109583342B (en) | Human face living body detection method based on transfer learning | |
Zhang et al. | Tv-gan: Generative adversarial network based thermal to visible face recognition | |
Bharadwaj et al. | Periocular biometrics: When iris recognition fails | |
CN105335722B (en) | Detection system and method based on depth image information | |
CN106934359B (en) | Multi-view gait recognition method and system based on high-order tensor subspace learning | |
KR102174595B1 (en) | System and method for identifying faces in unconstrained media | |
Perez et al. | Methodological improvement on local Gabor face recognition based on feature selection and enhanced Borda count | |
Ghoualmi et al. | An ear biometric system based on artificial bees and the scale invariant feature transform | |
WO2016145940A1 (en) | Face authentication method and device | |
CN105335732B (en) | Based on piecemeal and differentiate that Non-negative Matrix Factorization blocks face identification method | |
Fan et al. | Fusing dynamic deep learned features and handcrafted features for facial expression recognition | |
CN107392187B (en) | Face in-vivo detection method based on gradient direction histogram | |
Singh et al. | Gender and ethnicity classification of iris images using deep class-encoder | |
CN108446690B (en) | Human face in-vivo detection method based on multi-view dynamic features | |
Choi et al. | Shadow compensation in 2D images for face recognition | |
Arandjelović et al. | Achieving robust face recognition from video by combining a weak photometric model and a learnt generic face invariant | |
CN104408780A (en) | Face recognition attendance system | |
CN106407966B (en) | A kind of face identification method applied to attendance | |
CN108921215A (en) | A kind of Smoke Detection based on local extremum Symbiotic Model and energy spectrometer | |
Liu et al. | Iris recognition in visible spectrum based on multi-layer analogous convolution and collaborative representation | |
Liao et al. | Ensemble local fractional LDA for face recognition | |
CN103034840A (en) | Gender identification method | |
CN103235943A (en) | Principal component analysis-based (PCA-based) three-dimensional (3D) face recognition system | |
Karamizadeh et al. | Race classification using gaussian-based weight K-nn algorithm for face recognition | |
Matta et al. | Facial gender recognition using multiple sources of visual information |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181221 |