InfoBoost for selecting discriminative Gabor features

Li Bai, Linlin Shen

Research output: Chapter in Book/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

We proposed a novel boosting algorithm - InfoBoost. Though AdaBoost has been widely used for feature selection and classifier learning, many of the selected features are redundant. By incorporating mutual information into AdaBoost, InfoBoost fully examines the redundancy between candidate classifiers and selected classifiers. The classifiers thus selected are both accurate and non-redundant. Experimental results show that InfoBoost learned strong classifier has lower training error than AdaBoost. InfoBoost learning has also been applied to selecting discriminative Gabor features for face recognition. Even with the simple correlation distance measure and 1-NN classifier, the selected Gabor features achieve quite high recognition accuracy on the FERET database, where both expression and illumination variance are present. When only 140 features are used, InfoBoost selected features achieve 95.5% accuracy, about 2.5% higher than that achieved by AdaBoost.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages423-432
Number of pages10
DOIs
Publication statusPublished - 2005
Externally publishedYes
Event11th International Conference on Computer Analysis of Images and Patterns, CAIP 2005 - Versailles, France
Duration: 5 Sept 20058 Sept 2005

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3691 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference11th International Conference on Computer Analysis of Images and Patterns, CAIP 2005
Country/TerritoryFrance
CityVersailles
Period5/09/058/09/05

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'InfoBoost for selecting discriminative Gabor features'. Together they form a unique fingerprint.

Cite this