Abstract
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 language | English |
|---|---|
| Title of host publication | Advances in Visual Information Systems - 4th International Conference, VISUAL 2000, Proceedings |
| Editors | Robert Laurini |
| Publisher | Springer Verlag |
| Pages | 407-418 |
| Number of pages | 12 |
| ISBN (Print) | 3540411771 |
| DOIs | |
| State | Published - 1 Jan 2000 |
| Event | 4th International Conference on Visual Information Systems, VISUAL 2000 - Lyon, France Duration: 2 Nov 2000 → 4 Nov 2000 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 1929 |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Other
| Other | 4th International Conference on Visual Information Systems, VISUAL 2000 |
|---|---|
| Country | France |
| City | Lyon |
| Period | 2/11/00 → 4/11/00 |
Fingerprint
Cite this
}
Adaptive multi-class metric content-based image retrieval. / Peng, Jing.
Advances in Visual Information Systems - 4th International Conference, VISUAL 2000, Proceedings. ed. / Robert Laurini. Springer Verlag, 2000. p. 407-418 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1929).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Research › peer-review
TY - GEN
T1 - Adaptive multi-class metric content-based image retrieval
AU - Peng, Jing
PY - 2000/1/1
Y1 - 2000/1/1
N2 - 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.
AB - 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.
UR - http://www.scopus.com/inward/record.url?scp=84959050043&partnerID=8YFLogxK
U2 - 10.1007/3-540-40053-2_36
DO - 10.1007/3-540-40053-2_36
M3 - Conference contribution
SN - 3540411771
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 407
EP - 418
BT - Advances in Visual Information Systems - 4th International Conference, VISUAL 2000, Proceedings
A2 - Laurini, Robert
PB - Springer Verlag
ER -