A Simple feature generation method based on fisher linear discriminant analysis

Minoru Fukumi, Yasue Mitsukura

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

This paper presents a new iterative algorithm for feature generation, which is approximately derived based on geometrical interpretation of the Fisher linear discriminant (FLD) analysis. In a field of pattern recognition or signal processing, the principal component analysis (PCA) is popular for data compression and feature extraction. Furthermore, iterative learning algorithms for obtaining eigenvectors in PCA have been presented in a field of pattern recognition, image analysis, and neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLD analysis has been used in many fields, especially face image analysis. The drawback of FLD is a long computational time in compression of a large-sized between-class covariance and within-class covariance matrices. Usually FLD has to carry out minimization of a within-class variance. However in this case the inverse matrix of the within-class covariance matrix cannot be obtained, since data dimension is higher than the number of data and then it includes many zero eigenvalues. In order to overcome this difficulty, a new iterative feature generation method, a simple FLD is introduced and its effectiveness is demonstrated.

Original languageEnglish
Title of host publicationProceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005
EditorsM.W. Marcellin
Pages342-346
Number of pages5
Publication statusPublished - 2005
Externally publishedYes
EventSeventh IASTED International Conference on Signal and Image Processing, SIP 2005 - Honolulu, HI, United States
Duration: 2005 Aug 152005 Aug 17

Other

OtherSeventh IASTED International Conference on Signal and Image Processing, SIP 2005
CountryUnited States
CityHonolulu, HI
Period05/8/1505/8/17

Fingerprint

Discriminant analysis
Covariance matrix
Principal component analysis
Image analysis
Pattern recognition
Data compression
Eigenvalues and eigenfunctions
Learning algorithms
Feature extraction
Signal processing
Neural networks

Keywords

  • Face recognition
  • Feature generation
  • Fisher linear discriminant analysis
  • Pattern recognition
  • Principal component analysis

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Fukumi, M., & Mitsukura, Y. (2005). A Simple feature generation method based on fisher linear discriminant analysis. In M. W. Marcellin (Ed.), Proceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005 (pp. 342-346)

A Simple feature generation method based on fisher linear discriminant analysis. / Fukumi, Minoru; Mitsukura, Yasue.

Proceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005. ed. / M.W. Marcellin. 2005. p. 342-346.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Fukumi, M & Mitsukura, Y 2005, A Simple feature generation method based on fisher linear discriminant analysis. in MW Marcellin (ed.), Proceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005. pp. 342-346, Seventh IASTED International Conference on Signal and Image Processing, SIP 2005, Honolulu, HI, United States, 05/8/15.
Fukumi M, Mitsukura Y. A Simple feature generation method based on fisher linear discriminant analysis. In Marcellin MW, editor, Proceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005. 2005. p. 342-346
Fukumi, Minoru ; Mitsukura, Yasue. / A Simple feature generation method based on fisher linear discriminant analysis. Proceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005. editor / M.W. Marcellin. 2005. pp. 342-346
@inproceedings{955682cd59554ac18287ccb46d0c6773,
title = "A Simple feature generation method based on fisher linear discriminant analysis",
abstract = "This paper presents a new iterative algorithm for feature generation, which is approximately derived based on geometrical interpretation of the Fisher linear discriminant (FLD) analysis. In a field of pattern recognition or signal processing, the principal component analysis (PCA) is popular for data compression and feature extraction. Furthermore, iterative learning algorithms for obtaining eigenvectors in PCA have been presented in a field of pattern recognition, image analysis, and neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLD analysis has been used in many fields, especially face image analysis. The drawback of FLD is a long computational time in compression of a large-sized between-class covariance and within-class covariance matrices. Usually FLD has to carry out minimization of a within-class variance. However in this case the inverse matrix of the within-class covariance matrix cannot be obtained, since data dimension is higher than the number of data and then it includes many zero eigenvalues. In order to overcome this difficulty, a new iterative feature generation method, a simple FLD is introduced and its effectiveness is demonstrated.",
keywords = "Face recognition, Feature generation, Fisher linear discriminant analysis, Pattern recognition, Principal component analysis",
author = "Minoru Fukumi and Yasue Mitsukura",
year = "2005",
language = "English",
isbn = "0889865183",
pages = "342--346",
editor = "M.W. Marcellin",
booktitle = "Proceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005",

}

TY - GEN

T1 - A Simple feature generation method based on fisher linear discriminant analysis

AU - Fukumi, Minoru

AU - Mitsukura, Yasue

PY - 2005

Y1 - 2005

N2 - This paper presents a new iterative algorithm for feature generation, which is approximately derived based on geometrical interpretation of the Fisher linear discriminant (FLD) analysis. In a field of pattern recognition or signal processing, the principal component analysis (PCA) is popular for data compression and feature extraction. Furthermore, iterative learning algorithms for obtaining eigenvectors in PCA have been presented in a field of pattern recognition, image analysis, and neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLD analysis has been used in many fields, especially face image analysis. The drawback of FLD is a long computational time in compression of a large-sized between-class covariance and within-class covariance matrices. Usually FLD has to carry out minimization of a within-class variance. However in this case the inverse matrix of the within-class covariance matrix cannot be obtained, since data dimension is higher than the number of data and then it includes many zero eigenvalues. In order to overcome this difficulty, a new iterative feature generation method, a simple FLD is introduced and its effectiveness is demonstrated.

AB - This paper presents a new iterative algorithm for feature generation, which is approximately derived based on geometrical interpretation of the Fisher linear discriminant (FLD) analysis. In a field of pattern recognition or signal processing, the principal component analysis (PCA) is popular for data compression and feature extraction. Furthermore, iterative learning algorithms for obtaining eigenvectors in PCA have been presented in a field of pattern recognition, image analysis, and neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLD analysis has been used in many fields, especially face image analysis. The drawback of FLD is a long computational time in compression of a large-sized between-class covariance and within-class covariance matrices. Usually FLD has to carry out minimization of a within-class variance. However in this case the inverse matrix of the within-class covariance matrix cannot be obtained, since data dimension is higher than the number of data and then it includes many zero eigenvalues. In order to overcome this difficulty, a new iterative feature generation method, a simple FLD is introduced and its effectiveness is demonstrated.

KW - Face recognition

KW - Feature generation

KW - Fisher linear discriminant analysis

KW - Pattern recognition

KW - Principal component analysis

UR - http://www.scopus.com/inward/record.url?scp=33644548684&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33644548684&partnerID=8YFLogxK

M3 - Conference contribution

SN - 0889865183

SP - 342

EP - 346

BT - Proceedings of the Seventh IASTED International Conference on Signal and Image Processing, SIP 2005

A2 - Marcellin, M.W.

ER -