A new regularized linear discriminant analysis method to solve small sample size problems

CSpace

	A new regularized linear discriminant analysis method to solve small sample size problems
	Chen, WS ; Yuen, PC ; Huang, R
	2005-11-01
发表期刊	INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE
ISSN	0218-0014
卷号	19 期号:7 页码:917-935
摘要	This paper presents a new regularization technique to deal with the small sample size (S3) problem in linear discriminant analysis (LDA) based face recognition. Regularization on the within-class scatter matrix S-w has been shown to be a good direction for solving the S3 problem because the solution is found in full space instead of a subspace. The main limitation in regularization is that a very high computation is required to determine the optimal parameters. In view of this limitation, this paper re-defines the three-parameter regularization on the within-class scatter matrix S-omega(alpha beta gamma), which is suitable for parameter reduction. Based on the new definition of S-omega(alpha beta gamma), we derive a single parameter (t) explicit expression formula for determining the three parameters and develop a one-parameter regularization on the within-class scatter matrix. A simple and efficient method is developed to determine the value of t. It is also proven that the new regularized within-class scatter matrix S-omega(alpha beta gamma) approaches the original within-class scatter matrix S, as the single parameter tends to zero. A novel one-parameter regularization linear discriminant analysis (1PRLDA) algorithm is then developed. The proposed 1PRLDA method for face recognition has been evaluated with two public available databases, namely ORL and FERET databases. The average recognition accuracies of 50 runs for ORL and FERET databases are 96.65% and 94.00%, respectively. Comparing with existing LDA-based methods in solving the S3 problem, the proposed 1PRLDA method gives the best performance.
关键词	linear discriminant analysis small sample size problem face recognition
语种	英语
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000233686300005
出版者	WORLD SCIENTIFIC PUBL CO PTE LTD
引用统计
文献类型	期刊论文
条目标识符	http://ir.amss.ac.cn/handle/2S8OKBNM/1402
专题	中国科学院数学与系统科学研究院
通讯作者	Chen, WS
作者单位	1.Shenzhen Univ, Coll Sci, Shenzhen 518060, Peoples R China 2.Chinese Acad Sci, Key Lab Math Mechanizat, Beijing 100080, Peoples R China 3.Hong Kong Baptist Univ, Dept Comp Sci, Hong Kong, Hong Kong, Peoples R China 4.Zhongshan Sun Yat Sen Univ, Dept Comp Sci, Guangzhou, Guangdong, Peoples R China
推荐引用方式 GB/T 7714	Chen, WS,Yuen, PC,Huang, R. A new regularized linear discriminant analysis method to solve small sample size problems[J]. INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE,2005,19(7):917-935.
APA	Chen, WS,Yuen, PC,&Huang, R.(2005).A new regularized linear discriminant analysis method to solve small sample size problems.INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE,19(7),917-935.
MLA	Chen, WS,et al."A new regularized linear discriminant analysis method to solve small sample size problems".INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE 19.7(2005):917-935.