Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Face recognition method based on dictionary learning models

A dictionary learning and face recognition technology, applied in the field of face recognition based on dictionary learning model, can solve the problems of ignoring the irrelevance degree of the base signal, and the test data is difficult to generalize performance, etc., and achieve the effect of high accuracy

Inactive Publication Date: 2014-04-30
PEKING UNIV
View PDF2 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] If the base signal d in the dictionary D k If the dimension of the linearly stretched subspace is too different from the dimension of the space where the input signal y is located, using such a dictionary, even if it has a good reconstruction ability for the training data, it is difficult for the new test data. good generalization performance
The research results show that the degree of uncorrelation of the base signal in the dictionary is of great help to improve the accuracy of sparse representation reconstruction and the algorithm operation speed, but most of the sparse representation and dictionary learning models at this stage only focus on the reconstruction of the dictionary as a whole Performance and discriminative performance, ignoring the important factor of the uncorrelated degree of the base signal

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
  • Face recognition method based on dictionary learning models
  • Face recognition method based on dictionary learning models
  • Face recognition method based on dictionary learning models

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] Step 1: Map the training and testing face images to a low-dimensional space to obtain the training signal set matrix Y.

[0037] The input sample is the face sample picture in the Extended Yale B database, which contains a total of 2414 pictures of 38 people under different lighting conditions. All pictures have been standardized and the size is 168×192 pixels, such as figure 2shown. Randomly divide each person's sample into two parts, training sample and test sample, and stretch each sample picture into a vector, then normalize it into a unit vector, and then use PCA [4] Reduce all samples to a 504-dimensional space.

[0038] Step 2: Establish a dictionary learning model, input Y into the dictionary learning model, and obtain the dictionary D adapted to the training set, the sparse vector matrix X of the training set, and the linear classifier W. The dictionary learned in this embodiment contains 570 basic signals, and the sparse coefficient threshold T=16.

[0039...

Embodiment 2

[0091] This embodiment conducts experiments based on the CAS-PEAL-R1 face database. The CAS-PEAL-R1 face database contains 30,900 images of 1,040 individuals, including pose, expression, occlusion, and illumination changes. Independent experiments were carried out on these four data sets, and 7, 5, 6, 9 training samples and 1 testing sample were selected respectively. The number of sample categories in each experiment is 242, that is, there are 242 different people in the experimental images.

[0092] The operation steps are the same as those in the first embodiment. All samples are first normalized and dimensionally reduced to a 500-dimensional space. The learned dictionary contains 700 basic signals, and the sparse coefficient threshold is set to T=16.

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 invention discloses a face recognition method based on dictionary learning models. The method comprises the following steps of: mapping trained and tested face images to a low-dimension space to acquire a training signal set matrix; establishing the dictionary learning models which comprise an irrelevant dictionary learning (IDL) model and an unconstrained irrelevant dictionary learning (U-IDL) model; inputting the training signal set matrix into the IDL and U-IDL models, and solving the models to acquire an irrelevant dictionary and a linear classifier; acquiring a corresponding sparse vector of each picture belonging to a test sample based on the dictionary acquired in the last step by using a sparse expression algorithm; and inputting the sparse vectors into the linear classifier to acquire category labels of test sample pictures, wherein the result expressed by the category labels is used as the face recognition result. The invention provides the new models and the new method for dictionary learning problems in sparse expression, and the models and the method can be applied to mode identification and image classification problem under common conditions; and particularly, aiming at face recognition application, the dictionary learning method can achieve relatively high face recognition accuracy.

Description

technical field [0001] The invention belongs to the technical field of face recognition, and in particular relates to a face recognition method based on a dictionary learning model. Background technique [0002] Sparse representation technology based on overcomplete dictionary is a hot topic in the fields of computer vision, pattern recognition, and machine learning. It has been used in image denoising and patching, face recognition, image classification, video abnormal behavior detection, etc There have been many successful applications in the research field. [0003] Given a column vector consisting of K base signals An overcomplete dictionary composed of columns is a set of real numbers, n is the dimension of the base signal column vector, and K is the number of base signal column vectors), for the input signal column vector in n-dimensional space The sparse representation problem can be expressed as: [0004] min x ...

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): G06K9/00G06K9/66
Inventor 林通刘诗査红彬
Owner PEKING UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products