Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
PCA versus LDA
Autore:
Martinez, AM; Kak, AC;
Indirizzi:
Purdue Univ, Sch Elect & Comp Engn, Robot Vis Lab, W Lafayette, IN 47907 USA Purdue Univ W Lafayette IN USA 47907 t Vis Lab, W Lafayette, IN 47907 USA
Titolo Testata:
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
fascicolo: 2, volume: 23, anno: 2001,
pagine: 228 - 233
SICI:
0162-8828(200102)23:2<228:PVL>2.0.ZU;2-E
Fonte:
ISI
Lingua:
ENG
Soggetto:
HUMAN FACES; RECOGNITION; EIGENFACES;
Keywords:
face recognition; pattern recognition; principal components analysis; linear discriminant analysis; learning from undersampled distributions; small training data sets;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
Engineering, Computing & Technology
Citazioni:
19
Recensione:
Indirizzi per estratti:
Indirizzo: Martinez, AM Purdue Univ, Sch Elect & Comp Engn, Robot Vis Lab, W Lafayette, IN 47907 USA Purdue Univ W Lafayette IN USA 47907 Lafayette, IN 47907 USA
Citazione:
A.M. Martinez e A.C. Kak, "PCA versus LDA", IEEE PATT A, 23(2), 2001, pp. 228-233

Abstract

In the context of the appearance-based paradigm for object recognition, itis generally believed that algorithms based on LDA (Linear Discriminant Analysis) are superior to those based on PCA (Principal Components Analysis).in this communication, we show that this is not always the case. We present our case first by using intuitively plausible arguments and, then. by showing actual results on a face database. Our overall conclusion is that whenthe training data set is small, PCA can outperform LDA and, also, that PCAis less sensitive to different training data sets.

ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 23/01/20 alle ore 18:18:02