Mixture Subclass Discriminant Analysis

被引：46

作者：

Gkalelis, Nikolaos ^{[1
,2
]}

Mezaris, Vasileios ^{[1
]}

Kompatsiaris, Ioannis ^{[1
]}

机构：

[1] Ctr Res & Technol Hellas CERTH, Informat & Telemat Inst, Thermi 57001, Greece

[2] Univ London Imperial Coll Sci Technol & Med, Dept Elect & Elect Engn, London SW7 2AZ, England

来源：

IEEE SIGNAL PROCESSING LETTERS | 2011年 / 18卷 / 05期

关键词：

Discriminant analysis; event recognition; face recognition; feature extraction; Gaussian distribution; pattern recognition; RECOGNITION; REDUCTION;

D O I：

10.1109/LSP.2011.2127474

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

In this letter, mixture subclass discriminant analysis (MSDA) that alleviates two shortcomings of subclass discriminant analysis (SDA) is proposed. In particular, it is shown that for data with Gaussian homoscedastic subclass structure a) SDA does not guarantee to provide the discriminant subspace that minimizes the Bayes error, and, b) the sample covariance matrix can not be used as the minimization metric of the discriminant analysis stability criterion (DSC). Based on this analysis MSDA modifies the objective function of SDA and utilizes a novel partitioning procedure to aid discrimination of data with Gaussian homoscedastic subclass structure. Experimental results confirm the improved classification performance of MSDA.

引用

页码：319 / 322

页数：4

共 12 条

[1] Eigenfaces vs. Fisherfaces: Recognition using class specific linear projection [J].

Belhumeur, PN ;

Hespanha, JP ;

Kriegman, DJ .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1997, 19 (07) :711-720

[2]

Fukunaga K, 1990, INTRO STAT PATTERN R, V2nd

[3]

GKALELIS N, 2010, ACM MULTIMEDIA 2010

[4]

GRAHAM D., 1998, COMPUTER SYSTEMS SCI, V163, P446

[5]

Hastie T, 1996, J ROY STAT SOC B, V58, P155

[6] Multiclass linear dimension reduction by weighted pairwise Fisher criteria [J].

Loog, M ;

Duin, RPW ;

Haeb-Umbach, R .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2001, 23 (07) :762-766

[7] Fractional-step dimensionality reduction [J].

Lotlikar, R ;

Kothari, R .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2000, 22 (06) :623-627

[8] Regularization studies of linear discriminant analysis in small sample size scenarios with application to face recognition [J].

Lu, JW ;

Plataniotis, KN ;

Venetsanopoulos, AN .

PATTERN RECOGNITION LETTERS, 2005, 26 (02) :181-191

[9] Where are linear feature extraction methods applicable? [J].

Martínez, AM ;

Zhu, ML .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2005, 27 (12) :1934-1944

[10] Uncorrelated heteroscedastic LDA based on the weighted pairwise Chernoff criterion [J].

Qin, AK ;

Suganthan, PN ;

Loog, M .

PATTERN RECOGNITION, 2005, 38 (04) :613-616

← 1 2 →