A New LDA-based Face Recognition System Which Can Solve the Small Sample Size Problem

Li Fen Chen*, Hong Yuan Mark Liao, Ming Tat Ko, Chih-Ching Lin, Gwo Jong Yu

*Corresponding author for this work

Research output: Contribution to conferencePaperpeer-review

Abstract

A new LDA-based face recognition system is presented in this paper. Linear discriminant analysis (LDA) is one of the most popular linear projection techniques for feature, extraction. The major drawback of applying LDA is that it may encounter the small sample size problem. This problem arises whenever the number of samples is smaller than the dimensionality of samples. Under the circumstances, the sample scatter matrix may become singular and the execution of LDA may have computational difficulty. In this paper, we propose a new LDA-based technique which can solve the small sample size problem. We also prove that the most expressive vectors derived in the null space of the within-class scatter matrix using Principal Component Analysis (PCA) are equal to the optimal discriminant vectors derived in the original space using LDA. The experimental results showed that the new LDA process improves the performance of a face recognition system significantly.

Original languageEnglish
Pages282-286
Number of pages5
StatePublished - 1 Dec 1998
Event4th International Conference on Computer Science and Informatics, JCIS 1998, 1st International Workshop on High Performance, 1st International Workshop on Computer Vision, Pattern Recognition and Image Processing Volume 4 - Research Triangle Park, NC, United States
Duration: 23 Oct 199828 Oct 1998

Conference

Conference4th International Conference on Computer Science and Informatics, JCIS 1998, 1st International Workshop on High Performance, 1st International Workshop on Computer Vision, Pattern Recognition and Image Processing Volume 4
CountryUnited States
CityResearch Triangle Park, NC
Period23/10/9828/10/98

Fingerprint Dive into the research topics of 'A New LDA-based Face Recognition System Which Can Solve the Small Sample Size Problem'. Together they form a unique fingerprint.

Cite this