CN104392246B - It is a kind of based between class in class changes in faces dictionary single sample face recognition method - Google Patents
It is a kind of based between class in class changes in faces dictionary single sample face recognition method Download PDFInfo
- Publication number
- CN104392246B CN104392246B CN201410725260.5A CN201410725260A CN104392246B CN 104392246 B CN104392246 B CN 104392246B CN 201410725260 A CN201410725260 A CN 201410725260A CN 104392246 B CN104392246 B CN 104392246B
- Authority
- CN
- China
- Prior art keywords
- face
- class
- matrix
- mrow
- msubsup
- 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
- G06V40/172—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
- G06V40/161—Detection; Localisation; Normalisation
Abstract
The present invention propose it is a kind of based between class in class changes in faces dictionary single sample face recognition method, solves the problems, such as at present single sample face recognition algorithms limitation.Step 1：Obtain image expression of the facial image in compression domain；Step 2：Structure includes the facial image training sample matrix of k class；Step 3：Build face transformation matrices between the average face matrix of face database and class；Step 4：Low-rank and sparse constraint are added face transformation matrices class；Step 5：Similar matrix and interclass difference matrix between solution class；Step 6：Similar matrix and interclass difference matrix between average face matrix, class are projected to low dimensional space；Step 7：Similar matrix and interclass difference matrix between average face matrix after dimensionality reduction, class are normalized using method for normalizing, and the sparse coefficient vector based on facial image training sample matrix is solved using norm optimization algorithm iteration；Step 8：Column vector face label in the selection average face matrix corresponding with sparse coefficient maximum, the result as final recognition of face.
Description
Technical field
The invention belongs to computer vision and mode identification technology, is related to a kind of based on changes in faces word in class between class
Single sample face recognition method of allusion quotation.
Background technology
As a kind of important biometrics identification technology, it has naturality and the strong advantage of crypticity for recognition of face.
Therefore, extensive concern is caused in the past few decades, wherein having relatively simple but being relatively applied successfully special based on statistics
The face identification method of sign such as Eigenface, Fisherface, Laplacianfaces etc..In recent years, sparse expression is in face
Identification field is applied, and achieves larger success.Recognition of face sorting technique (Sparse based on sparse expression
Representation Classifier, SRC), its thought is using all training images as super complete dictionary, and is tested
Image can be expressed as the linear combination of a small number of facial images in dictionary.Wright et al. confirms that quick l can be passed through1Norm is optimal
Change algorithm to solve the problem, and effect best at present is achieved in face recognition experiment.At present, for sparse table
Insufficient present in intelligent's face recognition method, researcher proposes a variety of innovatory algorithms in succession.Such as add on the basis of SRC
Enter low-rank matrix recovery, face and noise section are distinguished using the irrelevance of dictionary expression.
But biological characteristic is used as using face so that recognition of face has congenital the shortcomings that being difficult to go beyond.First, although face
The similitude of structure positions beneficial to Face datection, but the differentiation to different faces brings inconvenience.Secondly, the profile of face is deposited
In lot of unstable, such as abundant human face expression, extraneous complex environment (illumination, imaging angle), the shelter of face
(mouth mask, sunglasses etc.) and the face's change brought with age.These unstability can all influence face recognition algorithms
Performance.In face recognition algorithms, the change between different faces changes between belonging to class, and the profile variation of identical face belongs to
Change in class.Change between class and in class make it that recognition of face is considered as living things feature recognition field even artificial intelligence neck
One of most difficult research topic in domain.And with the quick popularization of video monitoring, under many application scenarios, especially in big model
In the authentication occasion enclosed, such as law enforcement, driving license and the checking of passport card, we are typically only capable to adopt for everyone in database
Collect a sample image.In the case, must just be completed only according to the single image of face in different visual angles, illumination, screening
Recognition of face task under the changing factors such as gear, expression.But the recognition of face based on sparse expression that Wright et al. is proposed
Algorithm, it is desirable to provide everyone face face-image under various different change conditions.I.e. the algorithm needs substantial amounts of training
Sample database is to ensure the accuracy of algorithm.W.Deng et al. proposes extension sparse expression grader to solve the people of lack sampling
Face identifies problem, or even single sample recognition of face problem.Average face is introduced again subsequently, based on extension sparse expression grader adds people
Single sample face recognition algorithms of face change.
In recent years, the various signal processing methods based on compressive sensing theory, have become computer vision and pattern is known
One of other standard signal processing method.Therefore, it is also by field of face identification of the compressive sensing theory applied to sparse expression
Natural idea.A.Majumdar et al. confirms accidental projection to several graders being recently proposed, i.e., sparse grader
(SC), group SC and nearest neighbor classifier (NN) all have robustness.Because recognition of face needs to handle substantial amounts of high accuracy
Facial image, the concern of height is caused using compressed sensing progress dimensionality reduction in image area recognition of face.
The content of the invention
The defects of the invention aims to overcome prior art, solves the limitation of single sample face recognition algorithms at present
Property the problem of, propose it is a kind of based between class in class changes in faces dictionary single sample face recognition method.
The inventive method is achieved through the following technical solutions：
It is a kind of based between class in class changes in faces dictionary single sample face recognition method, comprise the following steps：
Step 1：Facial image in database is speculated into matrix progress projection mapping using random, facial image is obtained and exists
The image expression of compression domain；
Step 2：Structure includes the facial image training sample matrix of k class；
Step 3：Face transformation matrices between the average face matrix of face database and class are built, it is each in average face matrix
Row represent the average face of i-th of face training image.
Step 4：Low-rank and sparse constraint are added face transformation matrices class；
Step 5：Face transformation matrices class are further decomposed using augmentation Lagrange dictionary training algorithm, asked
Solve similar matrix and interclass difference matrix between class；
Step 6：Similar matrix and interclass difference matrix between average face matrix, class are projected to low using PCA dimension-reduction algorithms
Dimensional space；
Step 7：Similar matrix and interclass difference matrix between average face matrix after dimensionality reduction, class are entered using method for normalizing
Row normalized, and the sparse coefficient arrow based on facial image training sample matrix is solved using norm optimization algorithm iteration
Amount；
Step 8：Column vector face label in the selection average face matrix corresponding with sparse coefficient maximum, as most
The result of whole recognition of face.
Beneficial effects of the present invention：
The present invention can to influence the illumination in single sample face recognition algorithms, block, the noise such as attitudes vibration constructs
Noise dictionary, noise dictionary is divided into between class noise in noise and class, sparse table intelligent is based on so as to effectively raise
The robustness of face list specimen discerning algorithm.
Embodiment
The present invention is described in further details with reference to specific implementation example.Here, the schematic implementation of the present invention
Example and its illustrate to be used to explain the present invention, it is but not as a limitation of the invention.
The present invention propose it is a kind of based between class in class changes in faces dictionary single sample face recognition method, be built upon with
On the basis of machine projection and sparse representation theory, and for a kind of innovatory algorithm of single sample recognition of face problem proposition.This hair
The bright key for solving single sample recognition of face is to establish robust, the general noise model unrelated with facial image.This is made an uproar
Acoustic model is primarily directed to the expression shape change of face, ambient lighting change, human face posture change, face occluder (mouth mask, sunglasses
Deng) etc. be modeled to form dictionary.So as to which these noises be separated with face information, so as to improve single sample recognition of face
Accuracy rate.Its specific steps includes：
Step 1: facial image in database is speculated into matrix progress projection mapping using random, facial image is obtained
In the image expression of compression domain, y=Φ x are mathematically represented by.Wherein Φ is projection matrix, and x is facial image, and y is compression
The facial image in domain.
Wherein, for the training image and test image in face database, all using identical accidental projection matrix Φ
Carry out dimensionality reduction.
Step 2: structure includes the facial image training sample dictionary of k classWherein n=n1+n2...+nk, niFor the training sample of every a kind of face
Number.Each row in dictionary D matrixRepresented for the column vector of compression domain facial image.Arbitary inquiry or test sample
Y can be expressed as the linear combination of the dictionary, mathematically be represented by y=D α+e, and wherein e is random noise, and α is based on word
Allusion quotation D coefficient
Wherein, in recognition of face, n sample of k classes face individual is included.This n sample is formed into dictionary D, then word
Allusion quotation D each row represent a width compression domain facial image.
Step 3: the average face dictionary of structure face databaseFace changes dictionary between classEach row in average face dictionary P
PiRepresent the average face of i-th of face training image.
Wherein, each row represent the average value of all training image all pixels of a kind of face, class in average face dictionary
Between face change dictionary V each row represent training sample image and the difference of such face average face.
Step 4: changing dictionary V to face class adds low-rank and sparse constraint.
Wherein, low-rank optimization is that sparse constraint is various in order to constrain to extract the clean face in front from face
Face changes in faces.
Step 5: one is entered face change dictionary V class using augmentation Lagrange dictionary training algorithm according to formula (1)
Step is decomposed, and solves similar matrix E and interclass difference matrix G between class
Wherein, due to changing in face in dictionary, facial contour height correlation in every a kind of face, therefore can be by between class
Phase Sihe class inherited extracts different face changes.
Step 6: using PCA dimension-reduction algorithms by average face matrix, between class similar matrix E and interclass difference matrix G project to
Low dimensional space.
Step 7: similar matrix E and interclass difference matrix G between the average face matrix P after dimensionality reduction, class is used into L2Normalization
Method is normalized, and uses l according to formula (2)1-l2Norm optimization algorithm iteration solves is instructed based on facial image
Practice sparse coefficient vector C=[the α β γ] of sample dictionary.
Step 8: the column vector face label in the P matrixes corresponding with sparse coefficient α maximums is selected, as final
The result of recognition of face.
Wherein, previous step solves obtained sparse coefficient α and represents test image and the degree of correlation of training image, by piece
The coefficient for finding maximum is lifted, its corresponding face is exactly the face classification for the test image that algorithm obtains.
Since then, with regard to completing/realizing single sample recognition of face problem.
Above-described specific descriptions, the purpose, technical scheme and beneficial effect of invention are carried out further specifically
It is bright, it should be understood that the specific embodiment that the foregoing is only the present invention, the protection model being not intended to limit the present invention
Enclose, within the spirit and principles of the invention, any modification, equivalent substitution and improvements done etc., should be included in the present invention
Protection domain within.
Claims (2)
1. it is a kind of based between class in class changes in faces dictionary single sample face recognition method, it is characterised in that including following step
Suddenly：
Step 1：Facial image in database is speculated into matrix progress projection mapping using random, facial image is obtained and is compressing
The image expression in domain；
Step 2：Structure includes the facial image training sample matrix of k class；
Step 3：Face transformation matrices between the average face matrix of face database and class are built, each list in average face matrix
Show the average face of i-th of face training image；
Wherein, face transformation matrices V is specially between class：
<mrow>
<mi>V</mi>
<mo>=</mo>
<mo>&lsqb;</mo>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mn>1</mn>
</mrow>
<mn>1</mn>
</msubsup>
<mo>-</mo>
<msub>
<mi>P</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mn>1</mn>
</mrow>
<mn>2</mn>
</msubsup>
<mo>-</mo>
<msub>
<mi>P</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<mn>....</mn>
<mo>,</mo>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>i</mi>
</mrow>
<mn>1</mn>
</msubsup>
<mo>-</mo>
<msub>
<mi>P</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>i</mi>
</mrow>
<mn>2</mn>
</msubsup>
<mo>-</mo>
<msub>
<mi>P</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>k</mi>
</mrow>
<mn>1</mn>
</msubsup>
<mo>-</mo>
<msub>
<mi>P</mi>
<mi>k</mi>
</msub>
<mo>,</mo>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>k</mi>
</mrow>
<mn>2</mn>
</msubsup>
<mo>-</mo>
<msub>
<mi>P</mi>
<mi>k</mi>
</msub>
<mo>,</mo>
<mo>...</mo>
<mo>&rsqb;</mo>
</mrow>
Wherein,For compression domain facial image column vector represent, ni be per a kind of face training sample number, PiIt is average
Each row in face matrix, j=1,2, i=1,2 ..., k；
Step 4：Low-rank and sparse constraint are added face transformation matrices class；
Step 5：Face transformation matrices class are further decomposed using augmentation Lagrange dictionary training algorithm, solved
Similar matrix and interclass difference matrix between class；
Step 6：Similar matrix and interclass difference matrix between average face matrix, class are projected to low dimensional using PCA dimension-reduction algorithms
Space；
Step 7：Similar matrix E and interclass difference matrix G between average face matrix P after dimensionality reduction, class is entered using method for normalizing
Row normalized, and the sparse coefficient arrow based on facial image training sample matrix is solved using norm optimization algorithm iteration
Measure c=[α β γ]；
Step 8：Column vector face label in the selection average face matrix P corresponding with sparse coefficient α maximums, as final
The result of recognition of face.
2. it is as claimed in claim 1 it is a kind of based between class in class changes in faces dictionary single sample face recognition method, it is special
Sign is, more accurately modeling is carried out for a variety of face changes in faces, by these changes in faces and front face
Separate, so as to solve the problems, such as that single sample recognition of face septum reset changes the influence to accuracy of identification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410725260.5A CN104392246B (en) | 2014-12-03 | 2014-12-03 | It is a kind of based between class in class changes in faces dictionary single sample face recognition method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410725260.5A CN104392246B (en) | 2014-12-03 | 2014-12-03 | It is a kind of based between class in class changes in faces dictionary single sample face recognition method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104392246A CN104392246A (en) | 2015-03-04 |
| CN104392246B true CN104392246B (en) | 2018-02-16 |
Family
ID=52610147
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410725260.5A Active CN104392246B (en) | 2014-12-03 | 2014-12-03 | It is a kind of based between class in class changes in faces dictionary single sample face recognition method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104392246B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550634B (en) * | 2015-11-18 | 2019-05-03 | 广东微模式软件股份有限公司 | Human face posture recognition methods based on Gabor characteristic and dictionary learning |
| CN105844223A (en) * | 2016-03-18 | 2016-08-10 | 常州大学 | Face expression algorithm combining class characteristic dictionary learning and shared dictionary learning |
| CN106295609B (en) * | 2016-08-22 | 2019-05-10 | 河海大学 | Single sample face recognition method based on block sparsity structure low-rank representation |
| CN107330382A (en) * | 2017-06-16 | 2017-11-07 | 深圳大学 | The single sample face recognition method and device represented based on local convolution characteristic binding |
| CN107194378B (en) * | 2017-06-28 | 2020-11-17 | 深圳大学 | Face recognition method and device based on mixed dictionary learning |
| CN107766832A (en) * | 2017-10-30 | 2018-03-06 | 国网浙江省电力公司绍兴供电公司 | A kind of face identification method for field operation construction management |
| CN110232370B (en) * | 2019-06-21 | 2022-04-26 | 华北电力大学（保定） | Power transmission line aerial image hardware detection method for improving SSD model |
| CN110610148B (en) * | 2019-09-02 | 2022-02-08 | 南京邮电大学 | Privacy protection-oriented compressed sensing visual shielding video behavior identification method |
| CN111523404A (en) * | 2020-04-08 | 2020-08-11 | 华东师范大学 | Partial face recognition method based on convolutional neural network and sparse representation |
| CN111931665B (en) * | 2020-08-13 | 2023-02-21 | 重庆邮电大学 | Under-sampling face recognition method based on intra-class variation dictionary modeling |
| CN113158812B (en) * | 2021-03-25 | 2022-02-08 | 南京工程学院 | Single-sample face recognition method based on mixed expansion block dictionary sparse representation |
| CN114841294B (en) * | 2022-07-04 | 2022-10-28 | 杭州德适生物科技有限公司 | Classifier model training method and device for detecting chromosome structure abnormality |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102111418A (en) * | 2011-03-02 | 2011-06-29 | 北京工业大学 | Facial feature cryptographic key generation-based internet identity authentication method |
| CN103632138A (en) * | 2013-11-20 | 2014-03-12 | 南京信息工程大学 | Low-rank partitioning sparse representation human face identifying method |
| CN103996047A (en) * | 2014-03-04 | 2014-08-20 | 西安电子科技大学 | Hyperspectral image classification method based on compression spectrum clustering integration |
-
2014
- 2014-12-03 CN CN201410725260.5A patent/CN104392246B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102111418A (en) * | 2011-03-02 | 2011-06-29 | 北京工业大学 | Facial feature cryptographic key generation-based internet identity authentication method |
| CN103632138A (en) * | 2013-11-20 | 2014-03-12 | 南京信息工程大学 | Low-rank partitioning sparse representation human face identifying method |
| CN103996047A (en) * | 2014-03-04 | 2014-08-20 | 西安电子科技大学 | Hyperspectral image classification method based on compression spectrum clustering integration |
Non-Patent Citations (3)
| Title |
|---|
| 《Extended SRC: Undersampled Face Recognition via Intraclass Variant Dictionary》;Weihong Deng等;《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》;20120930;第23卷(第9期);全文 * |
| 《In Defense of Sparsity Based Face Recognition》;Weihong Deng等;《2013 IEEE Conference on Computer Vision and Pattern Recognition》;20131231;第399-406页 * |
| 《基于稀疏表征的单样本人脸识别》;畅雪萍等;《计算机工程》;20101130;第36卷(第21期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104392246A (en) | 2015-03-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104392246B (en) | It is a kind of based between class in class changes in faces dictionary single sample face recognition method | |
| CN107194341B (en) | Face recognition method and system based on fusion of Maxout multi-convolution neural network | |
| CN107145842B (en) | Face recognition method combining LBP characteristic graph and convolutional neural network | |
| CN106682598B (en) | Multi-pose face feature point detection method based on cascade regression | |
| CN106372581B (en) | Method for constructing and training face recognition feature extraction network | |
| CN106934359B (en) | Multi-view gait recognition method and system based on high-order tensor subspace learning | |
| CN105335732B (en) | Based on piecemeal and differentiate that Non-negative Matrix Factorization blocks face identification method | |
| CN102938065B (en) | Face feature extraction method and face identification method based on large-scale image data | |
| Li et al. | Overview of principal component analysis algorithm | |
| CN108182397B (en) | Multi-pose multi-scale human face verification method | |
| CN104008375B (en) | The integrated face identification method of feature based fusion | |
| CN101004791A (en) | Method for recognizing facial expression based on 2D partial least square method | |
| Zhu et al. | Logarithm gradient histogram: A general illumination invariant descriptor for face recognition | |
| CN106096517A (en) | A kind of face identification method based on low-rank matrix Yu eigenface | |
| CN107122725B (en) | Face recognition method and system based on joint sparse discriminant analysis | |
| CN109753864A (en) | A kind of face identification method based on caffe deep learning frame | |
| CN112464844A (en) | Human behavior and action recognition method based on deep learning and moving target detection | |
| Mohamed et al. | Avatar face recognition using wavelet transform and hierarchical multi-scale LBP | |
| CN104899604B (en) | A kind of feature level fusing method merged based on data set | |
| CN110188646B (en) | Human ear identification method based on fusion of gradient direction histogram and local binary pattern | |
| CN103745242A (en) | Cross-equipment biometric feature recognition method | |
| CN110969101A (en) | Face detection and tracking method based on HOG and feature descriptor | |
| CN103942545A (en) | Method and device for identifying faces based on bidirectional compressed data space dimension reduction | |
| CN103942572A (en) | Method and device for extracting facial expression features based on bidirectional compressed data space dimension reduction | |
| CN105389573B (en) | A kind of face identification method based on three value mode layering manufactures of part |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |