Feature generation by simple-FLDA for pattern recognition

M. Fukumi, S. Karungaru, Yasue Mitsukura

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

10 Citations (Scopus)

Abstract

In this paper, a new feature generation method for pattern recognition is proposed, which is approximately derived from geometrical interpretation of the Fisher linear discriminant analysis (FLDA). 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 such fields, including neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLDA has been used in many fields, especially face image analysis. The drawback of FLDA is a long computational time based on a large-sized covariance matrix and the issue that the within-class covariance matrix is usually singular. Generally FLDA 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 generally 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 FLDA is introduced and its effectiveness is demonstrated for pattern recognition problems.

Original languageEnglish
Title of host publicationProceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet
Pages730-734
Number of pages5
Volume2
Publication statusPublished - 2005
Externally publishedYes
EventInternational Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet Commerce, IAWTIC 2005 - Vienna, Austria
Duration: 2005 Nov 282005 Nov 30

Other

OtherInternational Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet Commerce, IAWTIC 2005
CountryAustria
CityVienna
Period05/11/2805/11/30

Fingerprint

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

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Fukumi, M., Karungaru, S., & Mitsukura, Y. (2005). Feature generation by simple-FLDA for pattern recognition. In Proceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet (Vol. 2, pp. 730-734). [1631555]

Feature generation by simple-FLDA for pattern recognition. / Fukumi, M.; Karungaru, S.; Mitsukura, Yasue.

Proceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet. Vol. 2 2005. p. 730-734 1631555.

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

Fukumi, M, Karungaru, S & Mitsukura, Y 2005, Feature generation by simple-FLDA for pattern recognition. in Proceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet. vol. 2, 1631555, pp. 730-734, International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet Commerce, IAWTIC 2005, Vienna, Austria, 05/11/28.
Fukumi M, Karungaru S, Mitsukura Y. Feature generation by simple-FLDA for pattern recognition. In Proceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet. Vol. 2. 2005. p. 730-734. 1631555
Fukumi, M. ; Karungaru, S. ; Mitsukura, Yasue. / Feature generation by simple-FLDA for pattern recognition. Proceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet. Vol. 2 2005. pp. 730-734
@inproceedings{9e43e97be39c41989c752a3bde62329b,
title = "Feature generation by simple-FLDA for pattern recognition",
abstract = "In this paper, a new feature generation method for pattern recognition is proposed, which is approximately derived from geometrical interpretation of the Fisher linear discriminant analysis (FLDA). 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 such fields, including neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLDA has been used in many fields, especially face image analysis. The drawback of FLDA is a long computational time based on a large-sized covariance matrix and the issue that the within-class covariance matrix is usually singular. Generally FLDA 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 generally 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 FLDA is introduced and its effectiveness is demonstrated for pattern recognition problems.",
author = "M. Fukumi and S. Karungaru and Yasue Mitsukura",
year = "2005",
language = "English",
isbn = "0769525040",
volume = "2",
pages = "730--734",
booktitle = "Proceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet",

}

TY - GEN

T1 - Feature generation by simple-FLDA for pattern recognition

AU - Fukumi, M.

AU - Karungaru, S.

AU - Mitsukura, Yasue

PY - 2005

Y1 - 2005

N2 - In this paper, a new feature generation method for pattern recognition is proposed, which is approximately derived from geometrical interpretation of the Fisher linear discriminant analysis (FLDA). 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 such fields, including neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLDA has been used in many fields, especially face image analysis. The drawback of FLDA is a long computational time based on a large-sized covariance matrix and the issue that the within-class covariance matrix is usually singular. Generally FLDA 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 generally 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 FLDA is introduced and its effectiveness is demonstrated for pattern recognition problems.

AB - In this paper, a new feature generation method for pattern recognition is proposed, which is approximately derived from geometrical interpretation of the Fisher linear discriminant analysis (FLDA). 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 such fields, including neural networks. Their effectiveness has been demonstrated in many applications. However, recently the FLDA has been used in many fields, especially face image analysis. The drawback of FLDA is a long computational time based on a large-sized covariance matrix and the issue that the within-class covariance matrix is usually singular. Generally FLDA 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 generally 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 FLDA is introduced and its effectiveness is demonstrated for pattern recognition problems.

UR - http://www.scopus.com/inward/record.url?scp=33847238419&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33847238419&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:33847238419

SN - 0769525040

SN - 9780769525044

VL - 2

SP - 730

EP - 734

BT - Proceedings - International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA 2005 and International Conference on Intelligent Agents, Web Technologies and Internet

ER -