CN107273817A - A kind of face identification method and system based on rarefaction representation and average Hash - Google Patents
A kind of face identification method and system based on rarefaction representation and average Hash Download PDFInfo
- Publication number
- CN107273817A CN107273817A CN201710379451.4A CN201710379451A CN107273817A CN 107273817 A CN107273817 A CN 107273817A CN 201710379451 A CN201710379451 A CN 201710379451A CN 107273817 A CN107273817 A CN 107273817A
- Authority
- CN
- China
- Prior art keywords
- face
- sample
- mrow
- test sample
- average
- 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.)
- Granted
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
- G06V40/168—Feature extraction; Face representation
- G06V40/169—Holistic features and representations, i.e. based on the facial image taken as a whole
-
- 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
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/513—Sparse representations
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of face identification method based on rarefaction representation and average Hash and system;Comprise the following steps:Face test sample and all people's face training sample are pre-processed:Colored human face sample is converted into gray level image, and normalizes face test sample and all people's face training sample;Openness feature between sample is extracted using sparse representation model, face test sample is encoded into the linear combination of all face training samples, sparse coefficient matrix of the face test sample on all face training samples is calculated;The spatial structure feature of face sample interior is extracted, face test sample and the average Hash feature of each face training sample is calculated;By the openness feature and the spatial structure Fusion Features of sample interior between face sample;The final generic of face test sample is judged using the reconstructed error between the average Hash feature of face test sample and the face test sample of reconstruct.
Description
Technical field
The present invention relates to a kind of face identification method based on rarefaction representation and average Hash and system.
Background technology
Recognition of face is one of core content in computer vision field, it be facial image is carried out feature extraction,
The process of Classification and Identification, after facial image feature is obtained, is handled by further Algorithm Analysis, is realized to facial image
Authentication.Recognition of face is a complicated synthesis, the technology such as fusion mode identification, computer vision, image procossing, can be used
In fields such as security protection inspection, security monitoring, criminal investigation tracking, identifications, it is with a wide range of applications.But due to by face
In the limitation of sample size, class between otherness, class the factors such as similitude, illumination variation influence, how from face sample
Notebook data, which is concentrated, to be efficiently extracted face characteristic and efficiently realizes face classification, so as to improve robustness and the people of recognition of face
The accuracy rate of face identification, is one of problem of field of face identification.Therefore, research is directed to small sample face database, realizes Shandong
The face recognition algorithms that rod is strong, discrimination is high are extremely important.
Rarefaction representation is proposed based on compressive sensing theory, and its main thought is that training sample is utilized in assorting process
To constitute dictionary, the rarefaction representation of test sample is realized by seeking linear combination of the test sample on dictionary, if test
Sample belongs to a certain classification, then the test sample can be by all training sample linear expressions of the category in theory.But
Rarefaction representation algorithm has taken into consideration only the openness and spatial structural form that have ignored sample interior between sample, works as facial image
When being influenceed by strong illumination variation, traditional rarefaction representation can reduce the robustness of recognition of face, face recognition accuracy rate
It can be decreased obviously.
The content of the invention
The purpose of the present invention is exactly that there is provided a kind of face based on rarefaction representation and average Hash in order to solve the above problems
Recognition methods and system, it effectively increases the robustness of face characteristic extraction, is conducive to improving face recognition accuracy rate, simultaneously
Improve the speed of recognition of face.
To achieve these goals, the present invention is adopted the following technical scheme that:
A kind of face identification method based on rarefaction representation and average Hash, comprises the following steps:
Step (1):Face test sample and all people's face training sample are pre-processed:Colored human face sample is turned
Change gray level image into, and normalize face test sample and all people's face training sample;
Step (2):Openness feature between sample is extracted using sparse representation model:Face test sample is encoded into institute
There is the linear combination of face training sample, calculate sparse coefficient matrix of the face test sample on all face training samples;
Step (3):Extract the spatial structure feature of face sample interior:Calculate face test sample and each face instruction
Practice the average Hash feature of sample;
Step (4):By the openness feature and the spatial structure Fusion Features of sample interior between face sample:Will be all
The average Hash feature of face training sample as new training sample, and obtained with new training sample and step (2) it is dilute
Sparse coefficient matrix reconstructs face test sample;
Step (5):Utilize the reconstruct between the average Hash feature of face test sample and the face test sample of reconstruct
Error judges the final generic of face test sample.
Step (1) the normalization face test sample and all face training samples are shown in formula (1) and (2):
Y=y/ | | y | |; (1)
Wherein, y ∈ RsColumn vector for face test sample represents that s=w × h, s is the size of gray level image, and w is ash
The length of image is spent, h is the width of gray level image;Represent that the column vector of j-th of face training sample of c classes is represented;R represents real
Number, | | y | | the mould of face test sample is represented,Represent the mould of face training sample.
The step (2) includes:
Step (21):Face test sample is encoded into the linear combination of all face training samples:
Y=XA=[X1,X2,...,Xn][A1,A2,...,An]T; (3)
Wherein, A=[A1,A2,...Ac...,An], A represents sparse system of all face training samples under rarefaction representation
Matrix number, AcWhat is represented is the sparse coefficient vector in c classes corresponding to all face training samples;
X=[X1,X2,...Xc...,Xn], X represents the column vector set of all face training samples, Xc∈Rs*mAndRs*mRepresent column vector set of the gray level image size for s m face training sample, XcTable
Show the column vector set of m face training sample of c classes.
Step (22):Sparse coefficient matrix is solved using object function and solution formula:
μ is the normal number of setting, and I is unit matrix;The solution of sparse coefficient matrix is represented, and
XTRepresent the transposition of the column vector of face training sample.
The step (3), including:
Step (31):Calculate the average pixel of each its own w × h pixel of face sample:
Wherein,Represent c classes jth (j=1,2 ..., m) p-th of pixel value of individual face training sample.
Step (32):The average pixel value of each pixel value of each face sample and the face sample is subjected to size ratio
Compared with:If pixel value is more than or equal to average pixel value, just two-value turns to 1 to pixel value, and pixel value is less than average pixel value, as
Just two-value turns to 0 to plain value, and specific formula is shown in formula (7) and formula (8):
IfThen
IfThen
Step (33):
Compared by the way that each pixel value and average pixel value are carried out into size, the average Hash that generation embodies face sample is special
Levy;
Average Hash character representation by i-th of face training sample of c classes isAnd be made up of 0 and 1, andH=[H1,H2,...,Hc,...,Hn];
It is y by the average Hash character representation of face test sampleh, and be made up of 0 and 1.HcRepresent the face instruction of c classes
Average Hash characteristic set, the H for practicing sample represent average Hash characteristic set, the H of all face training samplesnRepresent the n-th class
Face training sample average Hash characteristic set,Represent+1 face training sample of m (c-1) of c classes
Average Hash feature.
The step of step (4) is:
Face test sample is reconstructed using the sparse coefficient vector and such other new training sample of each classificationMeter
Calculating formula is:
The step of step (5) is:
Weigh face test sample average Hash feature yhWith the face test sample of reconstructBetween reconstructed error, if
The reconstructed error e of c classescMinimum, then judge that face test sample belongs to c classes, the formula for calculating reconstructed error is:
A kind of face identification system based on rarefaction representation and average Hash, including:
Pretreatment module:Face test sample and all people's face training sample are pre-processed:By colored human face sample
Originally gray level image is converted into, and normalizes face test sample and all people's face training sample;
Openness characteristic extracting module between sample:Using sparse representation model, face test sample is encoded into all
The linear combination of face training sample, calculates sparse coefficient matrix of the face test sample on all face training samples;
The spatial structure characteristic extracting module of face sample interior:Calculate face test sample and each face training sample
This average Hash feature;
Fusion Features module:By the openness feature and the spatial structure Fusion Features of sample interior between face sample:
Using the average Hash feature of all face training samples as new training sample, and obtained with new training sample and step (2)
To sparse coefficient matrix reconstruct face test sample;
Determination module:Utilize the reconstruct between the average Hash feature of face test sample and the face test sample of reconstruct
Error judges the final generic of face test sample.
Beneficial effects of the present invention:
1 face identification method is by the sample of openness feature between the sample of sparse representation model and average hash algorithm
Structural Characteristics are blended, for face Small Sample Database collection, are effectively overcome rarefaction representation algorithm and are only considered dilute between sample
Dredge property and have ignored the shortcoming of sample interior spatial structural form, while the redundancy in face feature vector can be removed,
Preferably extract the identification information of facial image.The robustness of face characteristic extraction is effectively increased, is conducive to improving face
Recognition accuracy, while improving the speed of recognition of face.
2 face identification methods to the change of strong illumination variation or contrast with very strong robustness, in brightness or
In the case that contrast changes, the average Hash feature of this method changes all without generation is significant, it is possible to prevente effectively from
Gamma correction is adjusted the influence brought, can then improve the recognition accuracy of face.
Brief description of the drawings
Fig. 1 is the flow chart of the inventive method.
Fig. 2 is the algorithm flow chart for calculating average Hash feature.
Embodiment
The invention will be further described with embodiment below in conjunction with the accompanying drawings.
The method of the invention is based on sparse representation model, using the space of average Hash feature extraction face sample interior
Structural information, by openness feature between the sample of sparse representation model and the sample inner structure feature phase of average hash algorithm
Merge to reconstruct face test sample, carry out the classification of classification to face test sample finally by reconstructed error.Idiographic flow
As shown in figure 1, comprising the following steps:
Step 1:Face test sample and all people's face training sample are pre-processed, i.e., turned colored human face sample
Change gray level image into, and normalize face test sample and all people's face training sample;
Wherein, the formula of normalization face test sample and all face training samples is as follows:
Y=y/ | | y | |
y∈RsFor face test sample, wherein s=w × h is the size of gray level image,Represent j-th of face of the i-th class
Training sample.
Step 2:In order to extract the openness feature between sample, using sparse representation model, face test sample is encoded
Into the linear combination of all face training samples, sparse coefficient matrix of the test sample on all face training samples is calculated;
The specific method for obtaining sparse coefficient matrix by sparse representation model includes:
1) face test sample is encoded into the linear combination of all face training samples, formula is as follows:
Y=XA=[X1,X2,...,Xn][A1,A2,...,An]T
A=[A1,A2,...,An], A represents sparse coefficient matrix of all face training samples under rarefaction representation, AjTable
What is shown is the sparse coefficient vector in jth class corresponding to all face training samples;X=[X1,X2,...,Xn], X is represented
The column vector set of all face training samples, Xc∈Rs*mRepresent m training sample set of c classes.
2) sparse coefficient matrix is solved using following object function and solution formula:
μ is less normal number, the unit matrix that I refers to.
Step 3:In order to extract the spatial structure feature of face sample interior, face test sample and each face are calculated
The average Hash feature of training sample;
The algorithm flow of average Hash feature is calculated as shown in Fig. 2 specifically including:
1) the average pixel of each its own w × h pixel of face sample is calculated, formula is as follows:
In above formula,Represent c classes jth (j=1,2 ..., the m) pth of individual face training sample
Individual pixel value.
2) each pixel value of each face sample is carried out into size with the average pixel value of this face sample to be compared:If picture
Element value is more than average pixel value, and just two-value turns to 1 to pixel value herein, and pixel value is less than average pixel value, herein pixel
Just two-value turns to 0 to value, and specific formula is as follows:
IfThen
IfThen
3) compared by the way that each pixel value and average pixel value are carried out into size, the average that generation embodies this face sample is breathed out
Uncommon feature.Wherein, it is by the average Hash character representation of i-th of face training sample of c classesAnd by 0 and 1 group
Into, andH=[H1,H2,...,Hn];It is y by the average Hash character representation of face test sampleh,
And be made up of 0 and 1.
Step 4:By the openness feature and the spatial structure Fusion Features of sample interior between face sample, by owner
The average Hash feature of face training sample is as new training sample, and the sparse system obtained with new training sample and step 2
Matrix number reconstructs face test sample;
Reconstruct face test sample specific method be:Utilize the sparse coefficient vector and such other new instruction of each classification
Practice sample to reconstruct face test sampleCalculation formula is:
Step 5:Utilize the reconstructed error between the average Hash feature and reconstruct face test sample of face test sample
To judge the final generic of face test sample.
Specific method is:
Weigh face test sample average Hash feature yhWith the face test sample of reconstructBetween reconstructed error, if
The reconstructed error e of c classescMinimum, then judge that face test sample belongs to c classes, the formula for calculating reconstructed error is:
Although above-mentioned the embodiment of the present invention is described with reference to accompanying drawing, not to present invention protection model
The limitation enclosed, one of ordinary skill in the art should be understood that on the basis of technical scheme those skilled in the art are not
Need to pay various modifications or deform still within protection scope of the present invention that creative work can make.
Claims (7)
1. a kind of face identification method based on rarefaction representation and average Hash, it is characterized in that, comprise the following steps:
Step (1):Face test sample and all people's face training sample are pre-processed:Colored human face sample is converted into
Gray level image, and normalize face test sample and all people's face training sample;
Step (2):Openness feature between sample is extracted using sparse representation model:Face test sample is encoded into owner
The linear combination of face training sample, calculates sparse coefficient matrix of the face test sample on all face training samples;
Step (3):Extract the spatial structure feature of face sample interior:Calculate face test sample and each face training sample
This average Hash feature;
Step (4):By the openness feature and the spatial structure Fusion Features of sample interior between face sample:By all faces
The average Hash feature of training sample is as new training sample, and the sparse system obtained with new training sample and step (2)
Matrix number reconstructs face test sample;
Step (5):Utilize the reconstructed error between the average Hash feature of face test sample and the face test sample of reconstruct
To judge the final generic of face test sample.
2. the method as described in claim 1, it is characterized in that, step (1) the normalization face test sample and all faces
Training sample is shown in formula (1) and (2):
Y=y/ | | y | |; (1)
<mrow>
<msubsup>
<mi>x</mi>
<mi>j</mi>
<mi>c</mi>
</msubsup>
<mo>=</mo>
<msubsup>
<mi>x</mi>
<mi>j</mi>
<mi>c</mi>
</msubsup>
<mo>/</mo>
<mo>|</mo>
<mo>|</mo>
<msubsup>
<mi>x</mi>
<mi>j</mi>
<mi>c</mi>
</msubsup>
<mo>|</mo>
<mo>|</mo>
<mo>;</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, y ∈ RsColumn vector for face test sample represents that s=w × h, s is the size of gray level image, and w is gray level image
Length, h be gray level image width;Represent that the column vector of j-th of face training sample of c classes is represented;R represents real number, | | y
| | the mould of face test sample is represented,Represent the mould of face training sample.
3. the method as described in claim 1, it is characterized in that, the step (2) includes:
Step (21):Face test sample is encoded into the linear combination of all face training samples:
Y=XA=[X1,X2,...,Xn][A1,A2,...,An]T; (3)
Wherein, A=[A1,A2,...Ac...,An], A represents sparse coefficient square of all face training samples under rarefaction representation
Battle array, AcWhat is represented is the sparse coefficient vector in c classes corresponding to all face training samples;
X=[X1,X2,...Xc...,Xn], X represents the column vector set of all face training samples, Xc∈Rs*mAndRs*mRepresent column vector set of the gray level image size for s m face training sample, XcTable
Show the column vector set of m face training sample of c classes;
Step (22):Sparse coefficient matrix is solved using object function and solution formula:
<mrow>
<mover>
<mi>A</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<mi>arg</mi>
<mi> </mi>
<msub>
<mi>min</mi>
<mi>A</mi>
</msub>
<mo>{</mo>
<mo>|</mo>
<mo>|</mo>
<mi>y</mi>
<mo>-</mo>
<mi>X</mi>
<mi>A</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mn>2</mn>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<mi>&mu;</mi>
<mo>|</mo>
<mo>|</mo>
<mi>A</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mn>2</mn>
<mn>2</mn>
</msubsup>
<mo>}</mo>
<mo>;</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mover>
<mi>A</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>X</mi>
<mi>T</mi>
</msup>
<mi>X</mi>
<mo>+</mo>
<mi>&mu;</mi>
<mi>I</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msup>
<mi>X</mi>
<mi>T</mi>
</msup>
<mi>y</mi>
<mo>;</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
μ is the normal number of setting, and I is unit matrix;The solution of sparse coefficient matrix is represented, and
XTRepresent the transposition of the column vector of face training sample.
4. the method as described in claim 1, it is characterized in that, the step (3), including:
Step (31):Calculate the average pixel of each its own w × h pixel of face sample:
<mrow>
<mover>
<mi>k</mi>
<mo>&OverBar;</mo>
</mover>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mi>w</mi>
<mo>&times;</mo>
<mi>h</mi>
</mrow>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>p</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>w</mi>
<mo>&times;</mo>
<mi>h</mi>
</mrow>
</munderover>
<msubsup>
<mi>k</mi>
<mi>p</mi>
<mrow>
<mi>c</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msubsup>
<mo>;</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
</mrow>
1
Wherein,Represent p-th of pixel value of j-th of face training sample of c classes;J=1,2 ..., m;
Step (32):Each pixel value of each face sample is carried out into size with the average pixel value of the face sample to be compared:
If pixel value is more than or equal to average pixel value, just two-value turns to 1 to pixel value, and pixel value is less than average pixel value, pixel
Just two-value turns to 0 to value, and specific formula is shown in formula (7) and formula (8):
Step (33):
Compared by the way that each pixel value and average pixel value are carried out into size, generation embodies the average Hash feature of face sample;
Average Hash character representation by i-th of face training sample of c classes isAnd be made up of 0 and 1, andH=[H1,H2,...,Hc,...,Hn];
It is y by the average Hash character representation of face test sampleh, and be made up of 0 and 1;HcRepresent the face training sample of c classes
This average Hash characteristic set, H represent average Hash characteristic set, the H of all face training samplesnRepresent the people of the n-th class
The average Hash characteristic set of face training sample,Represent the average of m (c-1)+1 face training sample of c classes
Hash feature.
5. the method as described in claim 1, it is characterized in that, it is the step of step (4):
Utilize the sparse coefficient vector of each classificationWith such other new training sample HcTo reconstruct face test sampleMeter
Calculating formula is:
<mrow>
<mover>
<mi>y</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<msub>
<mi>H</mi>
<mi>c</mi>
</msub>
<msub>
<mover>
<mi>A</mi>
<mo>^</mo>
</mover>
<mi>c</mi>
</msub>
<mo>;</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>9</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
6. method as claimed in claim 5, it is characterized in that, it is the step of step (5):
Weigh face test sample average Hash feature yhWith the face test sample of reconstructBetween reconstructed error, if c classes
Reconstructed error ecMinimum, then judge that face test sample belongs to c classes, the formula for calculating reconstructed error is:
<mrow>
<msub>
<mi>e</mi>
<mi>c</mi>
</msub>
<mo>=</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>y</mi>
<mi>h</mi>
</msub>
<mo>-</mo>
<msub>
<mi>H</mi>
<mi>c</mi>
</msub>
<msub>
<mover>
<mi>A</mi>
<mo>^</mo>
</mover>
<mi>c</mi>
</msub>
<mo>|</mo>
<msub>
<mo>|</mo>
<mn>2</mn>
</msub>
<mo>;</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>10</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
7. a kind of face identification system based on rarefaction representation and average Hash, it is characterized in that, including:
Pretreatment module:Face test sample and all people's face training sample are pre-processed:Colored human face sample is turned
Change gray level image into, and normalize face test sample and all people's face training sample;
Openness characteristic extracting module between sample:Using sparse representation model, face test sample is encoded into all faces
The linear combination of training sample, calculates sparse coefficient matrix of the face test sample on all face training samples;
The spatial structure characteristic extracting module of face sample interior:Calculate face test sample and each face training sample
Average Hash feature;
Fusion Features module:By the openness feature and the spatial structure Fusion Features of sample interior between face sample:By institute
There is the average Hash feature of face training sample as new training sample, and obtained with new training sample and step (2)
Sparse coefficient matrix reconstructs face test sample;
Determination module:Utilize the reconstructed error between the average Hash feature of face test sample and the face test sample of reconstruct
To judge the final generic of face test sample.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710379451.4A CN107273817B (en) | 2017-05-25 | 2017-05-25 | A kind of face identification method and system based on rarefaction representation and mean value Hash |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710379451.4A CN107273817B (en) | 2017-05-25 | 2017-05-25 | A kind of face identification method and system based on rarefaction representation and mean value Hash |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107273817A true CN107273817A (en) | 2017-10-20 |
CN107273817B CN107273817B (en) | 2019-09-13 |
Family
ID=60065394
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710379451.4A Active CN107273817B (en) | 2017-05-25 | 2017-05-25 | A kind of face identification method and system based on rarefaction representation and mean value Hash |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107273817B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112633399A (en) * | 2020-12-30 | 2021-04-09 | 郑州轻工业大学 | Sparse collaborative joint representation pattern recognition method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104463148A (en) * | 2014-12-31 | 2015-03-25 | 南京信息工程大学 | Human face recognition method based on image reconstruction and Hash algorithm |
CN105160312A (en) * | 2015-08-27 | 2015-12-16 | 南京信息工程大学 | Recommendation method for star face make up based on facial similarity match |
CN106250811A (en) * | 2016-06-15 | 2016-12-21 | 南京工程学院 | Unconfinement face identification method based on HOG feature rarefaction representation |
-
2017
- 2017-05-25 CN CN201710379451.4A patent/CN107273817B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104463148A (en) * | 2014-12-31 | 2015-03-25 | 南京信息工程大学 | Human face recognition method based on image reconstruction and Hash algorithm |
CN105160312A (en) * | 2015-08-27 | 2015-12-16 | 南京信息工程大学 | Recommendation method for star face make up based on facial similarity match |
CN106250811A (en) * | 2016-06-15 | 2016-12-21 | 南京工程学院 | Unconfinement face identification method based on HOG feature rarefaction representation |
Non-Patent Citations (3)
Title |
---|
LEI ZHANG 等: "Sparse Representation or Collaborative Representation: Which Helps Face Recognition", 《IEEE》 * |
M_ZHANGJB: "三种基于感知哈希算法的相似图像检索技术", 《CSDN博客》 * |
阚亮: "基于稀疏表示的人脸识别算法研究", 《万方数据》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112633399A (en) * | 2020-12-30 | 2021-04-09 | 郑州轻工业大学 | Sparse collaborative joint representation pattern recognition method |
Also Published As
Publication number | Publication date |
---|---|
CN107273817B (en) | 2019-09-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106611169B (en) | A kind of dangerous driving behavior real-time detection method based on deep learning | |
CN104866829B (en) | A kind of across age face verification method based on feature learning | |
CN103605972B (en) | Non-restricted environment face verification method based on block depth neural network | |
CN110399821B (en) | Customer satisfaction acquisition method based on facial expression recognition | |
CN111639558B (en) | Finger vein authentication method based on ArcFace Loss and improved residual error network | |
CN106651830A (en) | Image quality test method based on parallel convolutional neural network | |
CN110133610A (en) | ULTRA-WIDEBAND RADAR action identification method based on time-varying distance-Doppler figure | |
CN109255289B (en) | Cross-aging face recognition method based on unified generation model | |
CN108446601A (en) | A kind of face identification method based on sound Fusion Features | |
CN105117708A (en) | Facial expression recognition method and apparatus | |
CN112329721B (en) | Remote sensing small target detection method for model lightweight design | |
CN103246874B (en) | Face identification method based on JSM (joint sparsity model) and sparsity preserving projection | |
CN107798308B (en) | Face recognition method based on short video training method | |
CN106778512A (en) | Face identification method under the conditions of a kind of unrestricted based on LBP and depth school | |
CN114783034B (en) | Facial expression recognition method based on fusion of local sensitive features and global features | |
CN107767416A (en) | The recognition methods of pedestrian's direction in a kind of low-resolution image | |
CN109977887A (en) | A kind of face identification method of anti-age interference | |
CN105574489A (en) | Layered stack based violent group behavior detection method | |
CN109919921B (en) | Environmental impact degree modeling method based on generation countermeasure network | |
CN106529586A (en) | Image classification method based on supplemented text characteristic | |
CN112766283B (en) | Two-phase flow pattern identification method based on multi-scale convolution network | |
CN104463148B (en) | Face identification method based on Image Reconstruction and hash algorithm | |
CN114332942A (en) | Night infrared pedestrian detection method and system based on improved YOLOv3 | |
CN105868711A (en) | Method for identifying human body behaviors based on sparse and low rank | |
CN107832753A (en) | A kind of face feature extraction method based on four value weights and multiple classification |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |