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
- matrix
- class
- mrow
- inter
- 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
- 238000000034 method Methods 0.000 title claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims abstract description 62
- 238000012549 training Methods 0.000 claims abstract description 22
- 230000014509 gene expression Effects 0.000 claims abstract description 10
- 230000009467 reduction Effects 0.000 claims abstract description 7
- 238000005457 optimization Methods 0.000 claims abstract description 6
- 230000008859 change Effects 0.000 claims description 21
- 238000000513 principal component analysis Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 230000003190 augmentative effect Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 230000001815 facial effect Effects 0.000 abstract description 4
- 230000009466 transformation Effects 0.000 abstract 2
- 230000006835 compression Effects 0.000 abstract 1
- 238000007906 compression Methods 0.000 abstract 1
- 238000012360 testing method Methods 0.000 description 5
- 230000000875 corresponding effect Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000008921 facial expression Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Collating Specific Patterns (AREA)
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 the technical field of computer vision and pattern recognition, and relates to a single-sample face recognition method based on an intra-class face change dictionary.
Background
Face recognition, an important biometric feature recognition technology, has natureStrong sex and secrecy. Therefore, there has been a lot of attention in the past decades, and some of them are simpler but more successfully applied face recognition methods based on statistical features, such as Eigenface, Fisherface, laplacian face, etc. In recent years, sparse representation is applied to the field of face recognition, and has achieved great success. The idea of sparse representation-based face recognition classification (SRC) is to use all training images as an overcomplete dictionary, and the test image can be represented as a linear combination of a few face images in the dictionary. Wright et al demonstrated that the protein was produced by rapid1The norm optimization algorithm solves the problem and achieves the best effect in the face recognition experiment. At present, researchers have proposed a plurality of improved algorithms in succession aiming at the defects existing in the sparse expression face recognition method. For example, low rank matrix recovery is added on the basis of SRC, and the irrelevance of dictionary representation is utilized to distinguish human faces and noise parts.
However, the human face is taken as the biological feature, so that the human face recognition has the defect that the human face recognition is difficult to overcome in nature. Firstly, although the similarity of the face structure is beneficial to face detection and positioning, it is inconvenient to distinguish different faces. Secondly, there are many instabilities in the shape of the face, such as rich facial expressions, external complex environments (lighting, imaging angles, etc.), face masks (masks, sunglasses, etc.), and changes in the face with age. These instabilities can affect the performance of the face recognition algorithm. In the face recognition algorithm, the variation between different faces belongs to the variation between classes, and the appearance variation of the same face belongs to the variation within the class. Inter-class and intra-class variations make face recognition considered one of the most difficult research topics in the field of biometric recognition and even in the field of artificial intelligence. With the rapid spread of video surveillance, in many applications, especially in a wide range of authentication applications, such as law enforcement, driver's licenses, and passport card authentication, we can typically only collect one sample image for each individual in a database. In this case, the face recognition task under the varying factors of different viewing angles, illumination, occlusion, expression, etc. must be completed only according to a single image of the face. However, the face recognition algorithm based on sparse representation proposed by Wright et al needs to provide facial images of each person under various different changing conditions. Namely, the algorithm needs a large amount of training sample databases to ensure the accuracy of the algorithm. Deng et al propose to extend sparse representation classifiers to solve the undersampled face recognition problem, even the single-sample face recognition problem. Subsequently, a single-sample face recognition algorithm of average face plus face variation is introduced based on the extended sparse representation classifier.
In recent years, various signal processing methods based on the compressed sensing theory have become one of the standard signal processing methods for computer vision and pattern recognition. Therefore, applying the compressive sensing theory to the field of face recognition of sparse expression is also an idea of the sequential chapter. Majumdar et al demonstrate that stochastic projection is robust to several recently proposed classifiers, namely Sparse Classifier (SC), group SC and nearest neighbor classifier (NN). Because face recognition needs to process a large number of high-precision face images, high attention has been paid to dimension reduction by using compressed sensing in image domain face recognition.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, solve the problem of limitation of the current single-sample face recognition algorithm, and provide a single-sample face recognition method based on an intra-class face change dictionary.
The method is realized by the following technical scheme:
a single sample face recognition method based on an inter-class intra-class face change dictionary comprises the following steps:
step 1: carrying out projection mapping on the face image in the database by using a random speculative matrix to obtain the image expression of the face image in a compressed domain;
step 2: constructing a face image training sample matrix containing k classes;
and step 3: and constructing an average face matrix and an inter-class face change matrix of the face database, wherein each column in the average face matrix represents the average face of the ith personal face training image.
And 4, step 4: adding low rank and sparse constraint to the inter-class face change matrix;
and 5: further decomposing the inter-class face change matrix by adopting an augmented Lagrange dictionary training algorithm, and solving an inter-class similarity matrix and an inter-class difference matrix;
step 6: projecting the average face matrix, the inter-class similarity matrix and the inter-class difference matrix to a low-dimensional space by adopting a PCA (principal component analysis) dimension reduction algorithm;
and 7: carrying out normalization processing on the average face matrix, the inter-class similarity matrix and the inter-class difference matrix after dimension reduction by adopting a normalization method, and iteratively solving a sparse coefficient vector based on a face image training sample matrix by adopting a norm optimization algorithm;
and 8: and selecting the column vector face label in the average face matrix corresponding to the maximum value of the sparse coefficient as a final face recognition result.
The invention has the beneficial effects that:
the method can construct the noise dictionary for the noises which affect illumination, shielding, posture change and the like in the single-sample face recognition algorithm, and divides the noise dictionary into inter-class noises and intra-class noises, thereby effectively improving the robustness of the sparse expression face single-sample recognition algorithm.
Detailed Description
The present invention will be described in further detail with reference to specific examples. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
The invention provides a single-sample face recognition method based on an inter-class face change dictionary, which is an improved algorithm based on random projection and sparse representation theory and provided for the problem of single-sample face recognition. The key of the single-sample face recognition is to establish a robust and universal noise model irrelevant to a face image. The noise model mainly aims at the expression change of the face, the environmental illumination change, the face posture change, the face shelter (a mask, sunglasses and the like) and the like to form a dictionary through modeling. Therefore, the noise is separated from the face information, and the accuracy of single-sample face recognition is improved. The method comprises the following specific steps:
step one, carrying out projection mapping on the face image in the database by using a random projection matrix to obtain an image expression of the face image in a compressed domain, wherein the expression can be expressed as y ═ Φ x mathematically. Wherein phi is a projection matrix, x is a face image, and y is a face image in a compressed domain.
And for the training image and the test image in the face database, reducing the dimension by using the same random projection matrix phi.
Step two, constructing a face image training sample dictionary containing k classesWherein n is n1+ n2.. + nk niThe number of training samples of each type of human face. Each column in the dictionary D matrixAny query or test sample y may be represented as a linear combination of the dictionary, mathematically represented as y-D α + e, where e is random noise, and α is the dictionary D-based coefficients
The face recognition includes n samples of k-class individual faces. These n samples are grouped into a dictionary D, and each column of the dictionary D represents a compressed domain face image.
Step three, constructing an average face dictionary of the face databaseAnd between class face change dictionaryEach column P in the average face dictionary PiRepresenting the average face of the ith personal face training image.
Wherein each column in the average face dictionary represents an average value of all pixels of all training images of a class of face, and each column in the inter-class face variation dictionary V represents a difference value between a training sample image and the average face of the class of face.
And step four, adding low-rank and sparse constraints to the inter-class face change dictionary V.
The low-rank optimization is to extract a face with a clean front from the face, and the sparse constraint is to constrain various face changes.
Step five, adopting an augmented Lagrange dictionary training algorithm to further decompose the inter-class face change dictionary V according to the formula (1) to obtain an inter-class similarity matrix E and an inter-class difference matrix G
In the face change dictionary, the face contour of each type of face is highly correlated, so different face changes can be extracted through similarity and difference between classes.
And sixthly, projecting the average face matrix, the inter-class similarity matrix E and the inter-class difference matrix G to a low-dimensional space by adopting a PCA (principal component analysis) dimension reduction algorithm.
Seventhly, the average face matrix after dimension reductionP, inter-class similarity matrix E and inter-class difference matrix G adopt L2The normalization method is used for normalization processing and adopts l according to a formula (2)1-l2The norm optimization algorithm iteratively solves the sparse coefficient vector c ═ α β gamma based on the face image training sample dictionary]。
And step eight, selecting the column vector face label in the P matrix corresponding to the maximum value of the sparse coefficient α as the final face recognition result.
The sparse coefficient α obtained in the previous step represents the correlation degree between the test image and the training image, the maximum coefficient is found through enumeration, and the corresponding face is the face type of the test image obtained through the algorithm.
From this, the single sample face recognition problem is completed/achieved.
The above detailed description is intended to illustrate the objects, aspects and advantages of the present invention, and it should be understood that the above detailed description is only exemplary of the present invention and is not intended to limit the scope of the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (2)
1. A single sample face recognition method based on an inter-class face change dictionary is characterized by comprising the following steps:
step 1: carrying out projection mapping on the face image in the database by using a random speculative matrix to obtain the image expression of the face image in a compressed domain;
step 2: constructing a face image training sample matrix containing k classes;
and step 3: constructing an average face matrix and an inter-class face change matrix of a face database, wherein each column in the average face matrix represents an average face of an ith personal face training image;
the inter-class face change matrix V specifically includes:
<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 the column vector representation of the compressed domain face image, ni is the number of training samples of each class of face, PiFor each column in the average face matrix, j 1,2, i 1, 2.
And 4, step 4: adding low rank and sparse constraint to the inter-class face change matrix;
and 5: further decomposing the inter-class face change matrix by adopting an augmented Lagrange dictionary training algorithm, and solving an inter-class similarity matrix and an inter-class difference matrix;
step 6: projecting the average face matrix, the inter-class similarity matrix and the inter-class difference matrix to a low-dimensional space by adopting a PCA (principal component analysis) dimension reduction algorithm;
step 7, normalizing the average face matrix P, the inter-class similarity matrix E and the inter-class difference matrix G after dimensionality reduction by adopting a normalization method, and iteratively solving a sparse coefficient vector c based on a face image training sample matrix by adopting a norm optimization algorithm to be [ α β gamma ];
and 8, selecting the column vector face label in the average face matrix P corresponding to the maximum value of the sparse coefficient α as the final face recognition result.
2. The method as claimed in claim 1, wherein the face variation dictionary within the class is modeled more accurately for different face variations, and the face variations are separated from the front face, so as to solve the influence of the face variations on the recognition accuracy in the single sample face recognition problem.
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 | |
| Wu et al. | Multi-view low-rank dictionary learning for image classification | |
| Jing et al. | Multi-spectral low-rank structured dictionary learning for face recognition | |
| Hu | Enhanced gabor feature based classification using a regularized locally tensor discriminant model for multiview gait recognition | |
| CN103136516B (en) | The face identification method that visible ray and Near Infrared Information merge and system | |
| Xu et al. | Bimodal biometrics based on a representation and recognition approach | |
| CN107392107B (en) | Face feature extraction method based on heterogeneous tensor decomposition | |
| Shen et al. | Shape recognition by bag of skeleton-associated contour parts | |
| Qian et al. | Local structure-based image decomposition for feature extraction with applications to face recognition | |
| Lee et al. | Face image retrieval using sparse representation classifier with gabor-lbp histogram | |
| Zheng et al. | Improved sparse representation with low-rank representation for robust face recognition | |
| Zhao et al. | Two-dimensional color uncorrelated discriminant analysis for face recognition | |
| Nimbarte et al. | Age Invariant Face Recognition using Convolutional Neural Network. | |
| CN111723759B (en) | Unconstrained face recognition method based on weighted tensor sparse graph mapping | |
| CN111325275A (en) | Robust image classification method and device based on low-rank two-dimensional local discriminant map embedding | |
| Zhao et al. | Sparse tensor embedding based multispectral face recognition | |
| CN108108652B (en) | Cross-view human behavior recognition method and device based on dictionary learning | |
| Çevik et al. | A novel high-performance holistic descriptor for face retrieval | |
| Mitra | Gaussian mixture models for human face recognition under illumination variations | |
| CN110287973B (en) | Image feature extraction method based on low-rank robust linear discriminant analysis | |
| CN108304833A (en) | Face identification method based on MBLBP and DCT-BM2DPCA | |
| Feng et al. | A Fechner multiscale local descriptor for face recognition | |
| Arai et al. | Human gait gender classification using 2D discrete wavelet transforms energy | |
| Andrie et al. | A review of Chinese Academy of Sciences (CASIA) gait database as a human gait recognition dataset | |
| Yu et al. | Tensor discriminant analysis with multiscale features for action modeling and categorization |
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 |