A multi-class relevance feedback approach to image retrieval

Research output: Contribution to conferencePaperResearchpeer-review

1 Citation (Scopus)

Abstract

Relevance feedback methods for content-based image retrieval have shown promise in a variety of image database applications. These techniques assume two-class relevance feedback, relevant and irrelevant. While simple computationally, two-class relevance feedback often becomes inadequate in providing sufficient 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. For each given query, we estimate local feature relevance based on Chi-squared analysis using information provided by multi-class relevance feedback. Local feature relevance is then used to compute a flexible metric that is highly adaptive to query locations. As a result, local data distributions can be sufficiently exploited, whereby rapid performance improvement can be achieved. Experimental results using real image data validate the efficacy of our method.

Original languageEnglish
Pages46-49
Number of pages4
StatePublished - 1 Jan 2001
EventIEEE International Conference on Image Processing (ICIP) 2001 - Thessaloniki, Greece
Duration: 7 Oct 200110 Oct 2001

Other

OtherIEEE International Conference on Image Processing (ICIP) 2001
CountryGreece
CityThessaloniki
Period7/10/0110/10/01

Fingerprint

Image retrieval
Feedback
Information analysis

Cite this

Peng, J. (2001). A multi-class relevance feedback approach to image retrieval. 46-49. Paper presented at IEEE International Conference on Image Processing (ICIP) 2001, Thessaloniki, Greece.
Peng, Jing. / A multi-class relevance feedback approach to image retrieval. Paper presented at IEEE International Conference on Image Processing (ICIP) 2001, Thessaloniki, Greece.4 p.
@conference{e874945287ef4ac5b34725fc071d14c4,
title = "A multi-class relevance feedback approach to image retrieval",
abstract = "Relevance feedback methods for content-based image retrieval have shown promise in a variety of image database applications. These techniques assume two-class relevance feedback, relevant and irrelevant. While simple computationally, two-class relevance feedback often becomes inadequate in providing sufficient 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. For each given query, we estimate local feature relevance based on Chi-squared analysis using information provided by multi-class relevance feedback. Local feature relevance is then used to compute a flexible metric that is highly adaptive to query locations. As a result, local data distributions can be sufficiently exploited, whereby rapid performance improvement can be achieved. Experimental results using real image data validate the efficacy of our method.",
author = "Jing Peng",
year = "2001",
month = "1",
day = "1",
language = "English",
pages = "46--49",
note = "null ; Conference date: 07-10-2001 Through 10-10-2001",

}

Peng, J 2001, 'A multi-class relevance feedback approach to image retrieval' Paper presented at IEEE International Conference on Image Processing (ICIP) 2001, Thessaloniki, Greece, 7/10/01 - 10/10/01, pp. 46-49.

A multi-class relevance feedback approach to image retrieval. / Peng, Jing.

2001. 46-49 Paper presented at IEEE International Conference on Image Processing (ICIP) 2001, Thessaloniki, Greece.

Research output: Contribution to conferencePaperResearchpeer-review

TY - CONF

T1 - A multi-class relevance feedback approach to image retrieval

AU - Peng, Jing

PY - 2001/1/1

Y1 - 2001/1/1

N2 - Relevance feedback methods for content-based image retrieval have shown promise in a variety of image database applications. These techniques assume two-class relevance feedback, relevant and irrelevant. While simple computationally, two-class relevance feedback often becomes inadequate in providing sufficient 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. For each given query, we estimate local feature relevance based on Chi-squared analysis using information provided by multi-class relevance feedback. Local feature relevance is then used to compute a flexible metric that is highly adaptive to query locations. As a result, local data distributions can be sufficiently exploited, whereby rapid performance improvement can be achieved. Experimental results using real image data validate the efficacy of our method.

AB - Relevance feedback methods for content-based image retrieval have shown promise in a variety of image database applications. These techniques assume two-class relevance feedback, relevant and irrelevant. While simple computationally, two-class relevance feedback often becomes inadequate in providing sufficient 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. For each given query, we estimate local feature relevance based on Chi-squared analysis using information provided by multi-class relevance feedback. Local feature relevance is then used to compute a flexible metric that is highly adaptive to query locations. As a result, local data distributions can be sufficiently exploited, whereby rapid performance improvement can be achieved. Experimental results using real image data validate the efficacy of our method.

UR - http://www.scopus.com/inward/record.url?scp=0035168051&partnerID=8YFLogxK

M3 - Paper

SP - 46

EP - 49

ER -

Peng J. A multi-class relevance feedback approach to image retrieval. 2001. Paper presented at IEEE International Conference on Image Processing (ICIP) 2001, Thessaloniki, Greece.