Linear discriminant analysis for data with subcluster structure

Haesun Park, Jaegul Choo, Barry L. Drake, Jinwoo Kang

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

5 Citations (Scopus)

Abstract

Linear discriminant analysis (LDA) is a widely-used feature extraction method in classification. However, the original LDA has limitations due to the assumption of a unimodal structure for each cluster, which is sat- isfied in many applications such as facial image data when variations such as angle and illumination can significantly influence the images of the same person. In this paper, we propose a novel method, hierarchi- cal LDA(h-LDA), which takes into account hierarchical subcluster structures in the data sets. Our experiments show that regularized h-LDA produces better accuracy than LDA, PCA, and tensorFaces.

Original languageEnglish
Title of host publication2008 19th International Conference on Pattern Recognition, ICPR 2008
Publication statusPublished - 2008 Dec 1
Externally publishedYes
Event2008 19th International Conference on Pattern Recognition, ICPR 2008 - Tampa, FL, United States
Duration: 2008 Dec 82008 Dec 11

Publication series

NameProceedings - International Conference on Pattern Recognition
ISSN (Print)1051-4651

Other

Other2008 19th International Conference on Pattern Recognition, ICPR 2008
CountryUnited States
CityTampa, FL
Period08/12/808/12/11

Fingerprint

Discriminant analysis
Feature extraction
Lighting
Experiments

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Park, H., Choo, J., Drake, B. L., & Kang, J. (2008). Linear discriminant analysis for data with subcluster structure. In 2008 19th International Conference on Pattern Recognition, ICPR 2008 [4761084] (Proceedings - International Conference on Pattern Recognition).

Linear discriminant analysis for data with subcluster structure. / Park, Haesun; Choo, Jaegul; Drake, Barry L.; Kang, Jinwoo.

2008 19th International Conference on Pattern Recognition, ICPR 2008. 2008. 4761084 (Proceedings - International Conference on Pattern Recognition).

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

Park, H, Choo, J, Drake, BL & Kang, J 2008, Linear discriminant analysis for data with subcluster structure. in 2008 19th International Conference on Pattern Recognition, ICPR 2008., 4761084, Proceedings - International Conference on Pattern Recognition, 2008 19th International Conference on Pattern Recognition, ICPR 2008, Tampa, FL, United States, 08/12/8.
Park H, Choo J, Drake BL, Kang J. Linear discriminant analysis for data with subcluster structure. In 2008 19th International Conference on Pattern Recognition, ICPR 2008. 2008. 4761084. (Proceedings - International Conference on Pattern Recognition).
Park, Haesun ; Choo, Jaegul ; Drake, Barry L. ; Kang, Jinwoo. / Linear discriminant analysis for data with subcluster structure. 2008 19th International Conference on Pattern Recognition, ICPR 2008. 2008. (Proceedings - International Conference on Pattern Recognition).
@inproceedings{a6c92f145d1e42fa9dce509bf068018e,
title = "Linear discriminant analysis for data with subcluster structure",
abstract = "Linear discriminant analysis (LDA) is a widely-used feature extraction method in classification. However, the original LDA has limitations due to the assumption of a unimodal structure for each cluster, which is sat- isfied in many applications such as facial image data when variations such as angle and illumination can significantly influence the images of the same person. In this paper, we propose a novel method, hierarchi- cal LDA(h-LDA), which takes into account hierarchical subcluster structures in the data sets. Our experiments show that regularized h-LDA produces better accuracy than LDA, PCA, and tensorFaces.",
author = "Haesun Park and Jaegul Choo and Drake, {Barry L.} and Jinwoo Kang",
year = "2008",
month = "12",
day = "1",
language = "English",
isbn = "9781424421756",
series = "Proceedings - International Conference on Pattern Recognition",
booktitle = "2008 19th International Conference on Pattern Recognition, ICPR 2008",

}

TY - GEN

T1 - Linear discriminant analysis for data with subcluster structure

AU - Park, Haesun

AU - Choo, Jaegul

AU - Drake, Barry L.

AU - Kang, Jinwoo

PY - 2008/12/1

Y1 - 2008/12/1

N2 - Linear discriminant analysis (LDA) is a widely-used feature extraction method in classification. However, the original LDA has limitations due to the assumption of a unimodal structure for each cluster, which is sat- isfied in many applications such as facial image data when variations such as angle and illumination can significantly influence the images of the same person. In this paper, we propose a novel method, hierarchi- cal LDA(h-LDA), which takes into account hierarchical subcluster structures in the data sets. Our experiments show that regularized h-LDA produces better accuracy than LDA, PCA, and tensorFaces.

AB - Linear discriminant analysis (LDA) is a widely-used feature extraction method in classification. However, the original LDA has limitations due to the assumption of a unimodal structure for each cluster, which is sat- isfied in many applications such as facial image data when variations such as angle and illumination can significantly influence the images of the same person. In this paper, we propose a novel method, hierarchi- cal LDA(h-LDA), which takes into account hierarchical subcluster structures in the data sets. Our experiments show that regularized h-LDA produces better accuracy than LDA, PCA, and tensorFaces.

UR - http://www.scopus.com/inward/record.url?scp=77957962660&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77957962660&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:77957962660

SN - 9781424421756

T3 - Proceedings - International Conference on Pattern Recognition

BT - 2008 19th International Conference on Pattern Recognition, ICPR 2008

ER -