Pseudo relevance feedback based on iterative probabilistic one-class SVMs in web image retrieval

Jingrui He, Mingjing Li, Zhiwei Li, Hong Jiang Zhang, Hanghang Tong, Changshui Zhang

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

To improve the precision of top-ranked images returned by a web image search engine, we propose in this paper a novel pseudo relevance feedback method named iterative probabilistic one-class SVMs to re-rank the retrieved images. By assuming that most top-ranked images are relevant to the query, we iteratively train one-class SVMs, and convert the outputs to probabilities so as to combine the decision from different image representation. The effectiveness of our method is validated by systematic experiments even if the assumption is not well satisfied.

Original languageEnglish (US)
Pages (from-to)213-220
Number of pages8
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3332
StatePublished - 2004
Externally publishedYes

Fingerprint

Pseudo-relevance Feedback
Image retrieval
Image Retrieval
Search engines
Iterative methods
World Wide Web
Feedback
Search Engine
Experiments
Image Representation
Convert
Query
Class
Output
Experiment

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

@article{71b144b158d248eaaacce005b9099cc0,
title = "Pseudo relevance feedback based on iterative probabilistic one-class SVMs in web image retrieval",
abstract = "To improve the precision of top-ranked images returned by a web image search engine, we propose in this paper a novel pseudo relevance feedback method named iterative probabilistic one-class SVMs to re-rank the retrieved images. By assuming that most top-ranked images are relevant to the query, we iteratively train one-class SVMs, and convert the outputs to probabilities so as to combine the decision from different image representation. The effectiveness of our method is validated by systematic experiments even if the assumption is not well satisfied.",
author = "Jingrui He and Mingjing Li and Zhiwei Li and Zhang, {Hong Jiang} and Hanghang Tong and Changshui Zhang",
year = "2004",
language = "English (US)",
volume = "3332",
pages = "213--220",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Pseudo relevance feedback based on iterative probabilistic one-class SVMs in web image retrieval

AU - He, Jingrui

AU - Li, Mingjing

AU - Li, Zhiwei

AU - Zhang, Hong Jiang

AU - Tong, Hanghang

AU - Zhang, Changshui

PY - 2004

Y1 - 2004

N2 - To improve the precision of top-ranked images returned by a web image search engine, we propose in this paper a novel pseudo relevance feedback method named iterative probabilistic one-class SVMs to re-rank the retrieved images. By assuming that most top-ranked images are relevant to the query, we iteratively train one-class SVMs, and convert the outputs to probabilities so as to combine the decision from different image representation. The effectiveness of our method is validated by systematic experiments even if the assumption is not well satisfied.

AB - To improve the precision of top-ranked images returned by a web image search engine, we propose in this paper a novel pseudo relevance feedback method named iterative probabilistic one-class SVMs to re-rank the retrieved images. By assuming that most top-ranked images are relevant to the query, we iteratively train one-class SVMs, and convert the outputs to probabilities so as to combine the decision from different image representation. The effectiveness of our method is validated by systematic experiments even if the assumption is not well satisfied.

UR - http://www.scopus.com/inward/record.url?scp=35048812970&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35048812970&partnerID=8YFLogxK

M3 - Article

VL - 3332

SP - 213

EP - 220

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -