A Face Recognition Method Based on Sparse Coding

A face recognition and sparse coding technology, which is applied in the field of face recognition based on sparse coding, can solve problems such as occlusion or noise interference, difficulty in determining the true distribution of residuals, and poor face image quality. The effect of reducing computational complexity, improving recognition rate and robustness

Active Publication Date: 2019-10-01
SUN YAT SEN UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

When the face is occluded or any pixel is corrupted, the true distribution of the residual is difficult to determine
[0008] Although many face recognition methods have been proposed, due to the complexity of the face itself and the environment, face recognition technology still has many unsolved difficulties, especially when the face image quality is poor, occlusion or noise interference Time

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Face Recognition Method Based on Sparse Coding
  • A Face Recognition Method Based on Sparse Coding
  • A Face Recognition Method Based on Sparse Coding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0043] Such as figure 1 As shown, the method provided by the invention specifically includes the following steps:

[0044] The first step, assuming that there are k known objects in the training sample set, among which n objects contained in the i-th object i training samples are expressed as a matrix where i=1,2,...,k,v ij ∈ R m ,j=1,2,...,n i , v ij Represents the column vector corresponding to the j-th training sample of the i-th object, m represents v ij dimension, the training sample set A can be expressed as:

[0045] A=[A 1 ,A 2 ,...,A k ]=[v 11 ,v 12 ,...,v knk ]∈R m×n ,

[0046] in Indicates the total number of training samples;

[0047] In the second step, let the test sample be expressed as y, and let the reconstructed sample y rec Initialized to the mean of all training samples;

[0048] The third step, calculate y and y rec The residual between e=y-y rec ;

[0049] The fourth step is to define the diagonal matrix P=diag(p 1 ,p 2 ,...,p m );...

Embodiment 2

[0063] In order to prove the effectiveness and robustness of the method provided by the present invention, two groups of comparative experiments were carried out in this implementation. In this embodiment, the first set of comparative experiments is done on the Extended Yale-B face database. In the first set of experiments, Subset1 and Subset2 in the EYB face database were used as training sets, and there were a total of 717 training sample images, which were collected under suitable lighting conditions. Using Subset3 as the test set, there are a total of 453 test samples, and these images are all subject to strong lighting. The size of each image is 96×84.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The present invention relates to a face recognition method based on sparse coding. The traditional sparse coding method often presupposes that the coding residue satisfies a certain probability distribution form, such as Laplace distribution or Gaussian distribution. On this basis, a l 1 or l 2 normal form to solve for sparsely encoded coefficients. Such processing measures can seriously affect the robustness of face recognition in some cases, especially when there are occlusions, noise or other forms of interference. The method provided by the present invention aims at the inadequacies of the traditional sparse coding method and introduces the idea of ​​iterative optimization, which mainly solves the following two problems: first, it does not need to presuppose the distribution form of the residual, and avoids the unreasonable residual distribution function. The impact of the robustness of face recognition; the second is to selectively retain a part of the useful pixels for recognition, while greatly reducing the amount of calculation, it better solves the problems of occlusion and pixel damage, and obtains more Robust recognition performance.

Description

technical field [0001] The present invention relates to the field of face recognition, and more specifically, to a face recognition method based on sparse coding. Background technique [0002] With the rapid development of bioengineering technology and computer technology, people's demand for security identification and identity verification issues is getting higher and higher. Face recognition technology with the advantages of non-contact, non-intrusive, friendliness, and scalability stands out among a variety of biometric technologies, and has been widely used in video retrieval, public security monitoring, access control, identity recognition and other fields. With the further maturity of technology and the improvement of social recognition, face recognition technology will be applied in more fields. [0003] At present, the commonly used face recognition methods can be roughly divided into three types: face recognition methods based on 2D images, face recognition method...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/46G06K9/62
CPCG06V40/161G06V40/168G06V10/40G06V10/513G06F18/214
Inventor 郑慧诚连丽娜董佳羽
Owner SUN YAT SEN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products