Media.Player

Random Sampling for Face Recognition

Summary

LDA is a popular feature extraction technique for face recognition. However, it often suffers from the small sample size problem when dealing with the high dimensional face data. Fisherface and Null Space LDA (N-LDA) are two conventional approaches to address this problem. But in many cases, these LDA classifiers are overfitted to the training set and discard some useful discriminative information. In this paper, by analyzing different overfitting problems for the two kinds of LDA classifiers, we propose an approach using random subspace and bagging to improve them respectively. By random sampling on feature vector and training samples, multiple stabilized Fisherface and N-LDA classifiers are constructed. The two kinds of complementary classifiers are integrated using a fusion rule, so nearly all the discriminative information are preserved. We also apply this approach to the integration of multiple features. A robust face recognition system integrating shape, texture, and Gabor responses is finally developed.

Supplementary Information

Inventor: Tang, Xiao-ou
Priority Number: CN100356387C
IPC Current: G06K000900
Assignee Applicant: The Chinese University of Hong Kong
Title: Face recognition method based on random sampling | Based on face identifying method of random sampling
Usefulness: Face recognition method based on random sampling | Based on face identifying method of random sampling
Novelty: Face recognition method based on random sampling

Industry

ICT/Telecom

Sub Category

Image Processing

Application No.

04/ENG/182

Others

Inventor(s): Professor Xiaoou TANG, Department of Information Engineering

Licensing Status: Available for licensing

Patent Status:  Chinese Patent Pending  Hong Kong Patent Pending

Country/Region

Hong Kong