CN110046582A - Identify the color face recognition method of linear expression retaining projection based on multiple view - Google Patents

Identify the color face recognition method of linear expression retaining projection based on multiple view Download PDF

Info

Publication number
CN110046582A
CN110046582A CN201910313626.0A CN201910313626A CN110046582A CN 110046582 A CN110046582 A CN 110046582A CN 201910313626 A CN201910313626 A CN 201910313626A CN 110046582 A CN110046582 A CN 110046582A
Authority
CN
China
Prior art keywords
sample
training sample
linear expression
identified
multiple view
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
Application number
CN201910313626.0A
Other languages
Chinese (zh)
Other versions
CN110046582B (en
Inventor
刘茜
姜波
张佳垒
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Information Science and Technology
Original Assignee
Nanjing University of Information Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing University of Information Science and Technology filed Critical Nanjing University of Information Science and Technology
Priority to CN201910313626.0A priority Critical patent/CN110046582B/en
Publication of CN110046582A publication Critical patent/CN110046582A/en
Application granted granted Critical
Publication of CN110046582B publication Critical patent/CN110046582B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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
    • 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)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Human Computer Interaction (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses the color face recognition methods for identifying linear expression retaining projection based on multiple view, belong to technical field of face recognition, the present invention is to each color component images training sample, carry out the linear expression sample using other color component images training samples similar with its, obtains corresponding linear expression coefficient；The discriminatory analysis that all color component images training samples and its linear expression are carried out with multiple view, obtains optimum linearity projection vector；Training sample feature set and sample characteristics to be identified after being projected calculate sample characteristics to be identified to the distance of each training sample feature, sample to be identified are grouped into apart from the class where that the smallest training sample.Three chrominance components of colorized face images are considered as three views by this method, are made full use of the complementary information between the heterochromia information of colorized face images and different color, are effectively improved the effect of recognition of face.

Description

Identify the color face recognition method of linear expression retaining projection based on multiple view
Technical field
The invention belongs to technical field of face recognition, and in particular to identify the coloured silk of linear expression retaining projection based on multiple view Color face identification method.
Background technique
Number of patent application 201710800209.X is defined such as based on the face identification method for identifying linear expression retaining projection Under:
If X=[X1,X2,...,Xc] indicate the training sample set comprising c classification,It indicates i-th The training sample set of classification, XiInclude NiA sample, xij∈RdIndicate j-th of training sample of the i-th class, RdIndicate d dimension it is real to Duration set.
Pass through the following problem of solution first based on the face identification method for identifying linear expression retaining projection to be trained Sample xijLinear expression coefficient
Wherein, it enablesThen, optimum linearity projection vector v is obtained by solving following problem:
The data of single view can only be handled based on the face identification method for identifying linear expression retaining projection, it is therefore, right In colorized face images, the data (being usually to be converted into gray level image) of three chrominance components can only be first merged, are then reapplied This method.This way is ignored due to color difference and the complementary letter between the heterochromia information generated and different color Breath, to have an adverse effect to recognition of face effect.
Summary of the invention
Goal of the invention: the purpose of the present invention is to provide the colored human faces for identifying linear expression retaining projection based on multiple view Recognition methods makes full use of color difference using the data of multiple view learning art processing three chrominance components of colorized face images Complementary information between different information and different color can effectively improve the effect of recognition of face.
Technical solution: to achieve the above object, the present invention adopts the following technical scheme:
The color face recognition method for being identified linear expression retaining projection based on multiple view, is included the following steps:
Step 1, to each color component images training sample, other color component images training sample similar with its is used The original linear expression sample, obtains corresponding linear expression coefficient；
Step 2, the discriminatory analysis that all color component images training samples and its linear expression are carried out with multiple view, obtains Optimum linearity projection vector；
Step 3, training sample feature set and sample characteristics to be identified after being projected, calculate sample characteristics to be identified and arrive Sample to be identified is grouped into apart from the class where that the smallest training sample by the distance of each training sample feature.
Further, in the step 1, color component images training sample is obtained by solution formula (I)Line Property indicate coefficient
Wherein, X={ XR,XG,XBIndicate the colorized face images training sample set comprising tri- chrominance components of R, G, B,Indicate the i component image training sample set comprising c classification,Table Show XiIn p-th of classification training sample set,In include NpA training sample,Indicate XiIn p-th of classification Q training sample, i=R, G, B, j=R, G, B, p=1,2 ..., c, q=1,2 ..., Np, RdIndicate the real vector set of d dimension； It enables
Further, in the step 2, optimum linearity projection vector v is obtained by solution formula (II)R,vG,vB∈ Rd:
Wherein,
Further, in the step 3, specific practice is as follows:
Training sample feature set after projection is expressed as
Further, colorized face images sample y={ y to be identified for oneR,yG,yB, yi∈RdIndicate i points of y Measure image pattern, i=R, G, B；Sample characteristics to be identified after projection are expressed as
Calculate ZyTo the distance of each training sample feature, sample to be identified is grouped into apart from that the smallest training sample Class where this.
The utility model has the advantages that compared with prior art, the colour of the invention for identifying linear expression retaining projection based on multiple view Three chrominance components of colorized face images are considered as three views by face identification method, to each color component images training sample This use other training samples similar with its carry out the linear expression training sample, and to all color component images training samples And its linear expression carries out the discriminatory analysis of multiple view, makes full use of the heterochromia information and different color of colorized face images Between complementary information, effectively improve the effect of recognition of face.
Detailed description of the invention
Fig. 1 is face sample picture；
Fig. 2 is 20 random test discrimination wave patterns.
Specific embodiment
Below in conjunction with example and attached drawing, the present invention is described further.
The color face recognition method for being identified linear expression retaining projection based on multiple view, is included the following steps:
Step 1, to each color component images training sample, other color component images training sample similar with its is used The original linear expression sample, obtains corresponding linear expression coefficient；
Step 2, the discriminatory analysis that all color component images training samples and its linear expression are carried out with multiple view, obtains Optimum linearity projection vector；
Step 3, training sample feature set and sample characteristics to be identified after being projected, calculate sample characteristics to be identified and arrive Sample to be identified is grouped into apart from the class where that the smallest training sample by the distance of each training sample feature.
In step 1, color component images training sample is obtained by solution formula (I)Linear expression coefficient
Wherein, X={ XR,XG,XBIndicate the colorized face images training sample set comprising tri- chrominance components of R, G, B,Indicate the i component image training sample set comprising c classification,Table Show XiIn p-th of classification training sample set,In include NpA training sample,Indicate XiIn p-th of classification Q training sample, i=R, G, B, j=R, G, B, p=1,2 ..., c, q=1,2 ..., Np, RdIndicate the real vector set of d dimension； It enables
In step 2, optimum linearity projection vector v is obtained by solution formula (II)R,vG,vB∈Rd:
Wherein,
In step 3, specific practice is as follows:
Training sample feature set after projection is expressed as
Colorized face images sample y={ y to be identified for oneR,yG,yB, yi∈RdIndicate the i component image sample of y This, i=R, G, B.Sample characteristics to be identified after projection are expressed as
Calculate ZyTo the distance of each training sample feature, sample to be identified is grouped into apart from that the smallest training sample Class where this.
Face Recognition Grand Challenge (FRGC) version 2 is selected in experimental verification 4 colored human face database of Experiment (P.J.Phillips, P.J.Flynn, T.Scruggs, K.Bowyer, J.Chang, K.Hoffman,J.Marques,J.Min,W.Worek,“Overview of the Face Recognition Grand Challenge”,IEEE Conf.Computer Vision and Pattern Recognition,vol.1,pp.947- 954,2005).The database size is larger, contains tri- word banks of training, target, query, training word bank packet 12776 pictures containing 222 people, target word bank include 16028 pictures of 466 people, and query word bank includes 466 8014 pictures of people.100 people that experiment has selected training to gather, everyone 36 width images.All original graphs chosen It as all being corrected and (making two to be in a horizontal position), scales and cuts, each image pattern only retains 60 × 60 sizes Face and near zone.Treated, and face sample picture is shown in Fig. 1.
For based on the face identification method for identifying linear expression retaining projection, all images chosen in experimental data base Gray level image is all converted by original color image.In experimental data base, each classification randomly chooses 18 facial image samples This carries out 20 random tests as sample to be identified as training sample, remaining sample.
Fig. 2 is shown based on the face identification method (the DLRPP method i.e. in chart) for identifying linear expression retaining projection Identify color face recognition method (the MDLRPP method i.e. in chart) 20 times of linear expression retaining projection with based on multiple view The recognition effect of random test.In Fig. 2, abscissa is the serial number of random test, ordinate be discrimination (=correctly identify Number of samples to be identified/total sample number to be identified).Table 1 gives the discrimination mean value and mark of two methods, 20 random tests It is quasi- poor.Compared with based on the face identification method for identifying linear expression retaining projection, linear expression is identified based on multiple view and is retained The recognition effect of the color face recognition method of projection is significantly improved, this, which is demonstrated, identifies linear expression reservation based on multiple view The validity of the color face recognition method of projection.
The discrimination mean value and standard deviation and average workout times of 1 20 random tests of table
Method name Discrimination (mean value and standard deviation, %)
DLRPP 91.31±1.84
MDLRPP 93.41±2.01

Claims (5)

1. identifying the color face recognition method of linear expression retaining projection based on multiple view, which is characterized in that including walking as follows It is rapid:
Step 1, to each color component images training sample, come using other color component images training samples similar with its The linear expression sample obtains corresponding linear expression coefficient；
Step 2, the discriminatory analysis that all color component images training samples and its linear expression are carried out with multiple view, obtains optimal Linear projection vector；
Step 3, training sample feature set and sample characteristics to be identified after being projected calculate sample characteristics to be identified to each Sample to be identified is grouped into apart from the class where that the smallest training sample by the distance of a training sample feature.
2. the color face recognition method according to claim 1 for identifying linear expression retaining projection based on multiple view, It is characterized in that: in the step 1, color component images training sample being obtained by solution formula (I)Linear expression system Number
Wherein, X={ XR,XG,XBIndicate the colorized face images training sample set comprising tri- chrominance components of R, G, B,Indicate the i component image training sample set comprising c classification,Table Show XiIn p-th of classification training sample set,In include NpA training sample,Indicate XiIn p-th of classification Q training sample, i=R, G, B, j=R, G, B, p=1,2 ..., c, q=1,2 ..., Np, RdIndicate the real vector set of d dimension； It enables
3. the color face recognition method according to claim 2 for identifying linear expression retaining projection based on multiple view, It is characterized in that: in the step 2, optimum linearity projection vector v being obtained by solution formula (II)R,vG,vB∈Rd:
Wherein,
4. the color face recognition method according to claim 3 for identifying linear expression retaining projection based on multiple view, Be characterized in that: in the step 3, specific practice is as follows:
Training sample feature set after projection is expressed as
5. the color face recognition method according to claim 4 for identifying linear expression retaining projection based on multiple view, It is characterized in that: colorized face images sample y={ y to be identified for oneR,yG,yB, yi∈RdIndicate the i component image sample of y This, i=R, G, B；Sample characteristics to be identified after projection are expressed as
Calculate ZyTo the distance of each training sample feature, sample to be identified is grouped into apart from that the smallest training sample institute Class.
CN201910313626.0A 2019-04-18 2019-04-18 Color face recognition method based on multi-view discrimination linear representation preserving projection Active CN110046582B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910313626.0A CN110046582B (en) 2019-04-18 2019-04-18 Color face recognition method based on multi-view discrimination linear representation preserving projection

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910313626.0A CN110046582B (en) 2019-04-18 2019-04-18 Color face recognition method based on multi-view discrimination linear representation preserving projection

Publications (2)

Publication Number Publication Date
CN110046582A true CN110046582A (en) 2019-07-23
CN110046582B CN110046582B (en) 2020-06-02

Family

ID=67277702

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910313626.0A Active CN110046582B (en) 2019-04-18 2019-04-18 Color face recognition method based on multi-view discrimination linear representation preserving projection

Country Status (1)

Country Link
CN (1) CN110046582B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102708359A (en) * 2012-05-08 2012-10-03 北京工业大学 Face recognition method based on color images
CN106650769A (en) * 2016-09-28 2017-05-10 南京信息工程大学 Linear representation multi-view discrimination dictionary learning-based classification method
CN107392190A (en) * 2017-09-07 2017-11-24 南京信息工程大学 Color face recognition method based on semi-supervised multi views dictionary learning
CN107506744A (en) * 2017-09-07 2017-12-22 南京信息工程大学 Represent to retain based on local linear and differentiate embedded face identification method
CN107563334A (en) * 2017-09-07 2018-01-09 南京信息工程大学 Based on the face identification method for differentiating linear expression retaining projection

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102708359A (en) * 2012-05-08 2012-10-03 北京工业大学 Face recognition method based on color images
CN106650769A (en) * 2016-09-28 2017-05-10 南京信息工程大学 Linear representation multi-view discrimination dictionary learning-based classification method
CN107392190A (en) * 2017-09-07 2017-11-24 南京信息工程大学 Color face recognition method based on semi-supervised multi views dictionary learning
CN107506744A (en) * 2017-09-07 2017-12-22 南京信息工程大学 Represent to retain based on local linear and differentiate embedded face identification method
CN107563334A (en) * 2017-09-07 2018-01-09 南京信息工程大学 Based on the face identification method for differentiating linear expression retaining projection

Also Published As

Publication number Publication date
CN110046582B (en) 2020-06-02

Similar Documents

Publication Publication Date Title
CN104866829B (en) A kind of across age face verification method based on feature learning
CN107609493B (en) Method and device for optimizing human face image quality evaluation model
CN111325115B (en) Cross-modal countervailing pedestrian re-identification method and system with triple constraint loss
CN104063722B (en) A kind of detection of fusion HOG human body targets and the safety cap recognition methods of SVM classifier
CN104700078B (en) A kind of robot scene recognition methods based on scale invariant feature extreme learning machine
CN103839042B (en) Face identification method and face identification system
CN106897671B (en) Micro-expression recognition method based on optical flow and Fisher Vector coding
CN109902667A (en) Human face in-vivo detection method based on light stream guide features block and convolution GRU
CN111160189A (en) Deep neural network facial expression recognition method based on dynamic target training
CN106980809B (en) Human face characteristic point detection method based on ASM
CN109522883A (en) A kind of method for detecting human face, system, device and storage medium
CN104680545B (en) There is the detection method of well-marked target in optical imagery
CN111914761A (en) Thermal infrared face recognition method and system
CN105095867A (en) Rapid dynamic face extraction and identification method based deep learning
CN106599833B (en) A kind of face identification method adapted to based on field and manifold distance is measured
CN105740787B (en) Identify the face identification method of color space based on multicore
CN112364801A (en) Dynamic threshold face recognition method
CN115984930A (en) Micro expression recognition method and device and micro expression recognition model training method
CN109614927B (en) Micro expression recognition based on difference of front and rear frames and characteristic dimension reduction
CN111611909A (en) Multi-subspace-domain self-adaptive face recognition method
CN108038467B (en) A kind of sparse face identification method of mirror image in conjunction with thickness level
CN105184285A (en) Posture-spanning colored image facial expression recognition of direct push type migration group sparse discriminant analysis
CN108596906A (en) It is a kind of to refer to screen image quality evaluating method entirely based on sparse locality preserving projections
CN110046582A (en) Identify the color face recognition method of linear expression retaining projection based on multiple view
CN109766860A (en) Method for detecting human face based on improved Adaboost algorithm

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