CN108197529A - A kind of human facial feature extraction method for merging DLDP and sobel - Google Patents
A kind of human facial feature extraction method for merging DLDP and sobel Download PDFInfo
- Publication number
- CN108197529A CN108197529A CN201711208327.8A CN201711208327A CN108197529A CN 108197529 A CN108197529 A CN 108197529A CN 201711208327 A CN201711208327 A CN 201711208327A CN 108197529 A CN108197529 A CN 108197529A
- Authority
- CN
- China
- Prior art keywords
- dldp
- sobel
- sub
- image
- feature extraction
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/174—Facial expression recognition
- G06V40/175—Static expression
Abstract
A kind of human facial feature extraction method for merging DLDP and sobel is claimed in the present invention; 8 direction masks of Kirsch operators are divided into two sub- direction masks and calculate skirt response value respectively again by this method; two codings (DLDP1 and DLDP2) are obtained, the histogram for cascading two codings obtains expressive features DLDP.More useful information is obtained in order to enhance local feature, DLDP is merged with sobel operators to obtain feature Sobel DLDP.Experiment on JAFFE and Cohn Kanade Facial expression databases shows that this method compared with other feature extraction algorithms based on local grain, not only shortens the operation time of feature extraction, but also improves discrimination.
Description
Technical field
The invention belongs to image procossing and area of pattern recognition, particularly a kind of expressive features for merging DLDP and sobel
Extracting method.
Background technology
Facial expression recognition is used to analyze some emotions and intention of the mankind, in human-computer interaction, computer vision, Ren Leixin
There is important application in the fields such as theoretical and mood analog study of science.It is in facial expression recognition to extract effective expressive features
Committed step, it directly affects the discrimination of expression classification.Human facial feature extraction method is divided into based on global characteristics and is based on
The extracting method of Local textural feature.It is extracted compared to global characteristics, the feature extracting method based on part is to illumination, posture etc.
The robustness of variation is more preferable.For LBP since its principle is simple, computation complexity is low, is widely used in image identification, target detection
With tracking etc. fields.LBP is insensitive to consistent illumination variation, but the image table to incorporating non-uniform illumination variation and random noise
Show that effect is undesirable, influence best image recognition correct rate.For this purpose, Jabid in 2010 et al. propose it is a kind of to noise more
Local direction pattern (LDP) face of tool robustness describes method.2013, RiveraAR et al. was rung using the edge in 8 directions
The data in 2 directions of positive and negative directional response value maximum are encoded in should being worth, and enhancing discrimination is so as to improve discrimination.2015
RiveraAR et al. encodes to improve by the way that second largest strength information of the directional information of image and first is combined
Discrimination.2016, R.Srinivasa Perumal et al. to the LDP codes obtained in single piecemeal by carrying out X-ORing
Calculate the single code of each block, for the tight code DR-LDP of efficient face recognition, discrimination is improved for generation.But
It is studied more than being and is all directed to improve discrimination, do not consider to calculate the skirt response in 8 directions, while will also be to obtained
8 skirt responses take absolute value unifies sequence again, leads to entire expression characteristic extraction procedure than relatively time-consuming.
Invention content
Present invention seek to address that above problem of the prior art.It proposes a kind of time for shortening feature extraction and improves knowledge
The not human facial feature extraction method of the fusion DLDP and sobel of rate.Technical scheme is as follows:
A kind of human facial feature extraction method for merging DLDP and sobel, includes the following steps:
S1, Facial Expression Image and sobel (Sobel Operator) are subjected to convolution, the edge for obtaining Facial Expression Image is special
Reference ceases;
S2, the image that S1 is obtained is divided into nonoverlapping M × N number of subregion;
S3, make us the arbitrary pixel X of face face-image centered on point form a pros with 8, its field pixel
Shape, according to 8 direction masks of Kirsch operators, using four angles of square as the first sub- direction mask, with the side of square
Upper four midpoints pass through the first sub- direction mask and the second sub- direction mask and 8 fields respectively as the second sub- direction mask
Four angles of pixel value convolutional calculation and up and down the skirt response value in direction;
S4, the position where before ranking 3 skirt response value is set as 1, other positions are set as 0, according to image 4 sides
To response symbol directly carry out binary-coding, construct image local description, obtain DLDP1 and DLDP2 codings;
S5, the histogram that DLDP coding Grad is obtained by DLDP1 and DLDP2 cascade the Nogata of M × N number of subregion
Figure obtains the characteristic information of Facial Expression Image, then obtains the human facial feature extraction side of fusion DLDP and sobel by grader
The discrimination of method.
Further, Facial Expression Image and sobel operators are carried out convolution by the step S1, obtain Facial Expression Image
Edge feature, specifically include step:
Sobel operators and Facial Expression Image convolution are to obtain horizontal and vertical filter result:
By by IxAnd IyIt combines that edge feature is obtained:
Further, the step S3 is according to the mask set { M of 8 direction Kirsch operators0~M7}:
Wherein M1,M3,M5,M7As the first sub- direction mask, M0,M2,M4,M6As the second sub- direction mask, utilize simultaneously
Four angles of field pixel value convolutional calculation and upper bottom left of first sub- direction mask and the second sub- direction mask and 8 same directions
The skirt response value of right direction.Then binary-coding, construction image office are directly carried out according to response symbol of the image in 4 directions
Portion's description, is defined as:
Wherein,m1iRepresent i-th of skirt response value, m in the first sub- direction2iRepresent the second sub- direction
In i-th of skirt response value, mkIt is the big skirt response value of kth.
Further, the step S5 obtain DLDP decompose local direction pattern-coding Grad histogram specifically include
Step:
The DLDP codes of pixel form corresponding DLDP code patterns according to artwork coordinate arrangement, calculate each pixel (x, y)
After DLDP codes, histogram HDLDPiIt is defined as follows:
HDLDPi=∑x,yf(DLDPi(x,y),Ci)
In formula,CiRepresent the sobel-DLDP codes in the i-th sub- direction, i=1,2, the DLDP of gained
Description is the histogram of DLDP coding Grad, recycles following formula that the histogram of DLDP1 and DLDP2 is linked up acquisition
DLDP histograms HDLDP, as image feature representation.
Further, the human facial feature extraction computational methods of step 5) the fusion DLDP and sobel are:
The DLDP operators of combined level direction and vertical direction just obtain Sobel-DLDP operators:
Sobel-DLDP={ Sobel-DLDPx,Sobel-DLDPy}
Wherein,
Sobel-DLDPx={ Sobel-DLDP1x,Sobel-DLDP2x}
Sobel-DLDPy={ Sobel-DLDP1y,Sobel-DLDP2y}
DLDPxRepresent the horizontal direction component of DLDP codes, DLDPyRepresent the vertical direction component of DLDP codes, DLDP1xWith
DLDP2xThe horizontal component of the first sub- direction DLDP and the horizontal component of the second sub- direction DLDP, DLDP1 are represented respectivelyyWith
DLDP2yThe vertical component of the first sub- direction DLDP and the vertical component of the second sub- direction DLDP are represented respectively.
Extract the Sobel-DLDP histograms H of each sub-block respectively using following formulaSobel-DLDP。
In formula,CiRepresent the sobel-DLDP codes in i-th of sub- direction, i=1,2.Most at last
Facial expression image couples the feature histogram of each sub-block with sequence from left to right, from top to bottom, which is used for table
It lets others have a look at face expressive features.Finally classified using SVM to fuse information and identify human face expression.
Further, the step S1 first, is passed through by the use of JAFFE Facial expression databases as experimental data base
Sobel operators obtain the edge feature both horizontally and vertically of facial expression image, and the feature for then merging the two directions obtains
The edge feature of facial expression image;Data volume is reduced by binaryzation again;Then feature extraction is carried out by DLDP again and obtains face
Expressive features histogram carries out expression classification recognition for grader.
It advantages of the present invention and has the beneficial effect that:
The present invention is directed to the problem of LDP algorithm time loss is larger, it is proposed that a kind of innovatory algorithm, and-- decomposes part side
To pattern (DLDP), while in order to improve the Sobel-DLDP expressive features that discrimination proposes fusion DLDP and sobel operators
Extraction algorithm.Sobel-DLDP not only solves the problem of LDP extractions expressive features are slow, and preferably several with current effect
Local Feature Extraction is compared, and improves discrimination.It is carried out on JAFFE and Cohn-Kanade Facial expression databases real
It tests, demonstrates validity and superiority of this method in processing speed and discrimination.
Description of the drawings
Fig. 1 is the human facial feature extraction method flow based on fusion DLDP and sobel that the present invention provides preferred embodiment
Figure.
Fig. 2 is DLDP encoding examples.
Fig. 3 is Sobel-DLDP characteristic extraction procedures.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, detailed
Carefully describe.Described embodiment is only the part of the embodiment of the present invention.
The present invention solve above-mentioned technical problem technical solution be:
As shown in Figure 1, the present invention provides a kind of human facial feature extraction method for merging DLDP and sobel, feature exists
In including the following steps:
Facial Expression Image and sobel are carried out convolution, obtain the edge feature of Facial Expression Image by S1:
The step of seeking edge feature be:
Sobel operators and original image I convolution are to obtain horizontal and vertical filter result:
By by IxAnd IyIt combines that edge feature is obtained:
Facial Expression Image is divided into nonoverlapping M × N number of subregion by S2;
8 direction masks of Kirsch operators are divided into according to four angles and up and down two sub- direction masks, acquired by S3
Two four direction skirt response values:
8 direction masks of Kirsch operators are divided into according to four angles and up and down two sub- direction masks, enable face
Point centered on the arbitrary pixel X of image, while respectively four angles and up and down direction are calculated by two sub- masks
Skirt response value obtains two four direction skirt response values, i.e. m10、m11、m12、m13And m20、m21、m22、m23.Ask two four directions
To the step of skirt response value such as table 1 is:
Table 1
Binary-coding is directly carried out according to response symbol of the image in 4 directions, constructs image local description, definition
For:
Wherein,m1iRepresent i-th of skirt response value, m in the first sub- direction2iRepresent the second sub- direction
In i-th of skirt response value, mkIt is the big skirt response value of kth.
Position where before ranking 3 skirt response value is set as 1 by S4, and other positions are set as 0, obtain DLDP1 and
DLDP2 is encoded;
S5 obtains the histogram of DLDP coding Grad by DLDP1 and DLDP2, cascades the Nogata of M × N number of subregion
Figure obtains the characteristic information of Facial Expression Image, then obtains the human facial feature extraction side of fusion DLDP and sobel by grader
The discrimination of method.
The method for obtaining the histogram of DLDP coding Grad:
The DLDP codes of pixel form corresponding DLDP code patterns according to artwork coordinate arrangement.Calculate each pixel (x, y)
After DLDP codes, DLDP descriptions-- histogram HDLDPiIt is defined as follows:
HDLDPi=∑x,yf(DLDPi(x,y),Ci)
In formula,CiRepresent the sobel-DLDP codes in i-th of sub- direction, i=1,2.Gained
DLDP description are the histograms of DLDP coding Grad.DLDP describes details of the attached bag containing image, for example, edge, angle,
Point and other Local textural features.Can obtain the histogram of DLDP1 and DLDP2 by above formula, recycle following formula by DLDP1 and
The histogram of DLDP2, which links up, obtains DLDP histograms HDLDP, as image feature representation.
Finally, the human facial feature extraction computational methods of fusion DLDP and sobel are:
The DLDP operators of combined level direction and vertical direction just obtain Sobel-DLDP operators:
Sobel-DLDP={ Sobel-DLDPx,Sobel-DLDPy}
Wherein,
Sobel-DLDPx={ Sobel-DLDP1x,Sobel-DLDP2x}
Sobel-DLDPy={ Sobel-DLDP1y,Sobel-DLDP2y}
DLDPxRepresent the horizontal direction component of DLDP codes, DLDPyRepresent the vertical direction component of DLDP codes, DLDP1xWith
DLDP2xThe horizontal component of the first sub- direction DLDP and the horizontal component of the second sub- direction DLDP, DLDP1 are represented respectivelyyWith
DLDP2yThe vertical component of the first sub- direction DLDP and the vertical component of the second sub- direction DLDP are represented respectively.
Extract the Sobel-DLDP histograms H of each sub-block respectively using following formulaSobel-DLDP。
In formula,CiRepresent the sobel-DLDP codes in i-th of sub- direction, i=1,2.Most table at last
Feelings image couples the feature histogram of each sub-block with sequence from left to right, from top to bottom, which is used to represent
Human face expression feature.Finally classified using SVM to fuse information and identify human face expression.
As shown in Fig. 2, for DLDP encoding examples in the present invention.Image pixel is by four angles and is divided into two up and down
Sub- direction mask obtains two four direction skirt response values, the absolute value of eight skirt response values is ranked up, before ranking
Position where 3 skirt response value is set as 1, and other positions are set as 0, obtains DLDP1 and DLDP2 codings.
As shown in figure 3, for Sobel-DLDP characteristic extraction procedures in the present invention.Made using JAFFE Facial expression databases
For experimental data base.First, the edge feature both horizontally and vertically of facial expression image, Ran Hourong are obtained by sobel operators
The feature for closing the two directions obtains the edge feature of facial expression image;Data volume is further reduced by binaryzation again;Then again
Feature extraction is carried out by DLDP and obtains human face expression feature histogram, expression classification recognition is carried out for grader.
The above embodiment is interpreted as being merely to illustrate the present invention rather than limit the scope of the invention.
After the content for having read the record of the present invention, technical staff can make various changes or modifications the present invention, these equivalent changes
Change and modification equally falls into the scope of the claims in the present invention.
Claims (6)
- A kind of 1. human facial feature extraction method for merging DLDP and sobel, which is characterized in that include the following steps:S1, Facial Expression Image and sobel operators are subjected to convolution, obtain the edge feature information of Facial Expression Image;S2, the image that S1 is obtained is divided into nonoverlapping M × N number of subregion;S3, make us the arbitrary pixel X of face face-image centered on point form a square with 8, its field pixel, According to 8 direction masks of Kirsch operators, using four angles of square as the first sub- direction mask, on the side of square four A midpoint passes through the first sub- direction mask and the second sub- direction mask and 8 field pixels respectively as the second sub- direction mask It is worth four angles of convolutional calculation and up and down the skirt response value in direction;S4, the position where before ranking 3 skirt response value is set as 1, other positions are set as 0, according to image in 4 directions Response symbol directly carries out binary-coding, constructs image local description, obtains DLDP1 and DLDP2 codings;S5, the histogram that DLDP coding Grad is obtained by DLDP1 and DLDP2, cascade the histogram of M × N number of subregion, obtain The human facial feature extraction method of fusion DLDP and sobel is obtained to the characteristic information of Facial Expression Image, then by grader Discrimination.
- 2. the human facial feature extraction method of fusion DLDP and sobel according to claim 1, which is characterized in that the step Facial Expression Image and sobel operators are carried out convolution by rapid S1, are obtained the edge feature of Facial Expression Image, are specifically included step Suddenly:Sobel operators and Facial Expression Image convolution are to obtain horizontal and vertical filter result:By by IxAnd IyIt combines that edge feature is obtained:
- 3. the human facial feature extraction method of fusion DLDP and sobel according to claim 1, which is characterized in that the step Rapid S3 is according to the mask set { M of 8 direction Kirsch operators0~M7}:Wherein M1,M3,M5,M7As the first sub- direction mask M0,M2,M4,M6As the second sub- direction mask, while utilize first Four angles of field pixel value convolutional calculation of sub- direction mask and the second sub- direction mask and 8 same directions and up and down side To skirt response value, binary-coding is directly then carried out according to response symbol of the image in 4 directions, construction image local is retouched Son is stated, is defined as:Wherein,m1iRepresent i-th of skirt response value, m in the first sub- direction2iIt represents in the second sub- direction I skirt response value, mkIt is the big skirt response value of kth.
- 4. the human facial feature extraction method of fusion DLDP and sobel according to claim 1, which is characterized in that the step The histogram that rapid S5 obtains DLDP decomposition local direction pattern-coding Grad specifically includes step:The DLDP codes of pixel form corresponding DLDP code patterns according to artwork coordinate arrangement, calculate the DLDP of each pixel (x, y) After code, histogram HDLDPiIt is defined as follows:HDLDPi=∑x,yf(DLDPi(x,y),Ci)In formula,CiRepresent the sobel-DLDP codes in the i-th sub- direction, i=1,2, DLDP description of gained It is the histogram of DLDP coding Grad, following formula is recycled to link up the histogram of DLDP1 and DLDP2 and obtains DLDP Nogatas Scheme HDLDP, as image feature representation.
- 5. the human facial feature extraction method of fusion DLDP and sobel according to claim 4, which is characterized in that the step Rapid 5, fusion DLDP and sobel human facial feature extraction computational methods be:The DLDP operators of combined level direction and vertical direction just obtain Sobel-DLDP operators:Sobel-DLDP={ Sobel-DLDPx,Sobel-DLDPy}Wherein,Sobel-DLDPx={ Sobel-DLDP1x,Sobel-DLDP2x}Sobel-DLDPy={ Sobel-DLDP1y,Sobel-DLDP2y}DLDPxRepresent the horizontal direction component of DLDP codes, DLDPyRepresent the vertical direction component of DLDP codes, DLDP1xAnd DLDP2x The horizontal component of the first sub- direction DLDP and the horizontal component of the second sub- direction DLDP, DLDP1 are represented respectivelyyAnd DLDP2yRespectively Represent the vertical component of the first sub- direction DLDP and the vertical component of the second sub- direction DLDP;Extract the Sobel-DLDP histograms H of each sub-block respectively using following formulaSobel-DLDP。In formula,CiRepresent the sobel-DLDP codes in the i-th sub- direction, i=1,2.Most facial expression image at last Couple the feature histogram of each sub-block with sequence from left to right, from top to bottom, which is used to represent face table Feelings feature.Finally classified using SVM to fuse information and identify human face expression.
- 6. the human facial feature extraction method of fusion DLDP and sobel according to claim 5, which is characterized in that the step Rapid S1 first, the water of facial expression image is obtained by sobel operators by the use of JAFFE Facial expression databases as experimental data base The edge feature of gentle vertical direction, the feature for then merging the two directions obtain the edge feature of facial expression image;Pass through again Binaryzation reduces data volume;Then feature extraction is carried out by DLDP again and obtains human face expression feature histogram, for grader Carry out expression classification recognition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711208327.8A CN108197529B (en) | 2017-11-27 | 2017-11-27 | Expression feature extraction method integrating DLDP and sobel |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711208327.8A CN108197529B (en) | 2017-11-27 | 2017-11-27 | Expression feature extraction method integrating DLDP and sobel |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108197529A true CN108197529A (en) | 2018-06-22 |
CN108197529B CN108197529B (en) | 2021-09-10 |
Family
ID=62573201
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711208327.8A Active CN108197529B (en) | 2017-11-27 | 2017-11-27 | Expression feature extraction method integrating DLDP and sobel |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108197529B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109086692A (en) * | 2018-07-16 | 2018-12-25 | 北京建筑大学 | A kind of face identification device and method |
CN110348284A (en) * | 2019-05-20 | 2019-10-18 | 天津科技大学 | Face recognition features' extracting method is blocked based on local direction texture pattern |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105205460A (en) * | 2015-09-17 | 2015-12-30 | 电子科技大学 | Face expression feature extraction and recognition method based on maximum direction encoding |
CN105893941A (en) * | 2016-03-28 | 2016-08-24 | 电子科技大学 | Facial expression identifying method based on regional images |
CN106778529A (en) * | 2016-11-25 | 2017-05-31 | 南京理工大学 | A kind of face identification method based on improvement LDP |
-
2017
- 2017-11-27 CN CN201711208327.8A patent/CN108197529B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105205460A (en) * | 2015-09-17 | 2015-12-30 | 电子科技大学 | Face expression feature extraction and recognition method based on maximum direction encoding |
CN105893941A (en) * | 2016-03-28 | 2016-08-24 | 电子科技大学 | Facial expression identifying method based on regional images |
CN106778529A (en) * | 2016-11-25 | 2017-05-31 | 南京理工大学 | A kind of face identification method based on improvement LDP |
Non-Patent Citations (3)
Title |
---|
ZHANG YI ETC.: ""Facial expression recognition based on fused Feature of PCA and LDP"", 《IMAGE PROCESSING AND PATTERN RECOGNITION》 * |
王绎博: ""基于改进LDP特征的人脸识别方法"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
罗元等: ""一种改进的LDP面部表情特征提取方法"", 《光电技术应用》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109086692A (en) * | 2018-07-16 | 2018-12-25 | 北京建筑大学 | A kind of face identification device and method |
CN110348284A (en) * | 2019-05-20 | 2019-10-18 | 天津科技大学 | Face recognition features' extracting method is blocked based on local direction texture pattern |
Also Published As
Publication number | Publication date |
---|---|
CN108197529B (en) | 2021-09-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ahmed et al. | Automated facial expression recognition using gradient-based ternary texture patterns | |
Li et al. | A co-saliency model of image pairs | |
CN104346619B (en) | Living body feature vector extraction device and the vectorial extracting method of living body feature | |
Li et al. | Principal line-based alignment refinement for palmprint recognition | |
CN107516316B (en) | Method for segmenting static human body image by introducing focusing mechanism into FCN | |
CN103996052B (en) | Three-dimensional face gender classification method based on three-dimensional point cloud | |
CN105069447B (en) | A kind of recognition methods of human face expression | |
CN107463917B (en) | Improved LTP and two-dimensional bidirectional PCA fusion-based face feature extraction method | |
Mohamed et al. | An improved LBP algorithm for avatar face recognition | |
CN105956579A (en) | Rapid finger vein identification method integrating fuzzy template and point characteristics | |
He et al. | Automatic magnetic resonance image prostate segmentation based on adaptive feature learning probability boosting tree initialization and CNN-ASM refinement | |
CN106874942B (en) | Regular expression semantic-based target model rapid construction method | |
Olague et al. | Evolving head tracking routines with brain programming | |
Sharif et al. | Intelligent human action recognition: A framework of optimal features selection based on Euclidean distance and strong correlation | |
CN107992807A (en) | A kind of face identification method and device based on CNN models | |
Yu et al. | Spectral embedding based facial expression recognition with multiple features | |
Ibragimov et al. | Accurate landmark-based segmentation by incorporating landmark misdetections | |
Al-Juboori et al. | Palm vein verification using multiple features and locality preserving projections | |
CN112733627A (en) | Finger vein identification method based on fusion of local feature network and global feature network | |
CN108197529A (en) | A kind of human facial feature extraction method for merging DLDP and sobel | |
Iqbal et al. | Facial expression recognition with active local shape pattern and learned-size block representations | |
Yang et al. | Metric learning based object recognition and retrieval | |
CN112784722B (en) | Behavior identification method based on YOLOv3 and bag-of-words model | |
Mohamed et al. | Face recognition based on Wavelet transform and adaptive local binary pattern | |
Ebrahimian et al. | Automated person identification from hand images using hierarchical vision transformer network |
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 |