Adaptive multi-class metric content-based image retrieval

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


Relevance feedback methods for content-based image retrieval have shown promise in a variety of image database applications. These techniques assume two (relevant and irrelevant) class relevance feedback. While simple computationally, two class relevance feedback often becomes inadequate in providing suÆcient information to help rapidly improve retrieval performance. We propose a locally adaptive technique for content-based image retrieval that enables relevance feedback to take on multi-class form. We estimate a exible multi-class metric for computing retrievals based on Chi-squared distance analysis. As a result, local data distributions can be suÆciently exploited, whereby rapid performance improvement can be achieved. The eÆcacy of our method is validated and compared against other competing techniques using a number of real world data sets.

Original languageEnglish
Title of host publicationAdvances in Visual Information Systems - 4th International Conference, VISUAL 2000, Proceedings
EditorsRobert Laurini
PublisherSpringer Verlag
Number of pages12
ISBN (Print)3540411771
StatePublished - 2000
Event4th International Conference on Visual Information Systems, VISUAL 2000 - Lyon, France
Duration: 2 Nov 20004 Nov 2000

Publication series

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


Other4th International Conference on Visual Information Systems, VISUAL 2000


Dive into the research topics of 'Adaptive multi-class metric content-based image retrieval'. Together they form a unique fingerprint.

Cite this