Independent feature analysis for image retrieval

Jing Peng, Bir Bhanu

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

Abstract

Content-based image retrieval methods based on the Euclidean metric expect the feature space to be isotropic. They suffer from unequal differential relevance of features in computing the similarity between images in the input feature space. We propose a learning method that attempts to overcome this limitation by capturing local differential relevance of features based on user feedback. This feedback, in the form of accept or reject examples generated in response to a query image, is used to locally estimate the strength of features along each dimension while taking into consideration the correlation between features. This results in local neighborhoods that are constricted along feature dimensions that are most relevant, while enlongated along less relevant ones. In addition to exploring and exploiting local principal information, the system seeks a global space for efficient independent feature analysis by combining such local information. We provide experimental results that demonstrate the efficacy of our technique using real-world data.

Original languageEnglish
Title of host publicationMachine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings
EditorsMaria Petrou, Petra Perner
PublisherSpringer Verlag
Pages103-115
Number of pages13
ISBN (Print)3540665994, 9783540665991
StatePublished - 1 Jan 1999
Event1st International Workshop on Machine Learning and Data Mining in Pattern Recognition, MLDM 1999 - Leipzig, Germany
Duration: 16 Sep 199918 Sep 1999

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1715
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other1st International Workshop on Machine Learning and Data Mining in Pattern Recognition, MLDM 1999
CountryGermany
CityLeipzig
Period16/09/9918/09/99

Fingerprint

Image retrieval
Image Retrieval
Feedback
Feature Space
Content-based Image Retrieval
Unequal
Efficacy
Euclidean
Query
Metric
Computing
Experimental Results
Estimate
Demonstrate
Relevance

Cite this

Peng, J., & Bhanu, B. (1999). Independent feature analysis for image retrieval. In M. Petrou, & P. Perner (Eds.), Machine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings (pp. 103-115). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1715). Springer Verlag.
Peng, Jing ; Bhanu, Bir. / Independent feature analysis for image retrieval. Machine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings. editor / Maria Petrou ; Petra Perner. Springer Verlag, 1999. pp. 103-115 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{4f64107445d147b5a799362e35b05d80,
title = "Independent feature analysis for image retrieval",
abstract = "Content-based image retrieval methods based on the Euclidean metric expect the feature space to be isotropic. They suffer from unequal differential relevance of features in computing the similarity between images in the input feature space. We propose a learning method that attempts to overcome this limitation by capturing local differential relevance of features based on user feedback. This feedback, in the form of accept or reject examples generated in response to a query image, is used to locally estimate the strength of features along each dimension while taking into consideration the correlation between features. This results in local neighborhoods that are constricted along feature dimensions that are most relevant, while enlongated along less relevant ones. In addition to exploring and exploiting local principal information, the system seeks a global space for efficient independent feature analysis by combining such local information. We provide experimental results that demonstrate the efficacy of our technique using real-world data.",
author = "Jing Peng and Bir Bhanu",
year = "1999",
month = "1",
day = "1",
language = "English",
isbn = "3540665994",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "103--115",
editor = "Maria Petrou and Petra Perner",
booktitle = "Machine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings",

}

Peng, J & Bhanu, B 1999, Independent feature analysis for image retrieval. in M Petrou & P Perner (eds), Machine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1715, Springer Verlag, pp. 103-115, 1st International Workshop on Machine Learning and Data Mining in Pattern Recognition, MLDM 1999, Leipzig, Germany, 16/09/99.

Independent feature analysis for image retrieval. / Peng, Jing; Bhanu, Bir.

Machine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings. ed. / Maria Petrou; Petra Perner. Springer Verlag, 1999. p. 103-115 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1715).

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

TY - GEN

T1 - Independent feature analysis for image retrieval

AU - Peng, Jing

AU - Bhanu, Bir

PY - 1999/1/1

Y1 - 1999/1/1

N2 - Content-based image retrieval methods based on the Euclidean metric expect the feature space to be isotropic. They suffer from unequal differential relevance of features in computing the similarity between images in the input feature space. We propose a learning method that attempts to overcome this limitation by capturing local differential relevance of features based on user feedback. This feedback, in the form of accept or reject examples generated in response to a query image, is used to locally estimate the strength of features along each dimension while taking into consideration the correlation between features. This results in local neighborhoods that are constricted along feature dimensions that are most relevant, while enlongated along less relevant ones. In addition to exploring and exploiting local principal information, the system seeks a global space for efficient independent feature analysis by combining such local information. We provide experimental results that demonstrate the efficacy of our technique using real-world data.

AB - Content-based image retrieval methods based on the Euclidean metric expect the feature space to be isotropic. They suffer from unequal differential relevance of features in computing the similarity between images in the input feature space. We propose a learning method that attempts to overcome this limitation by capturing local differential relevance of features based on user feedback. This feedback, in the form of accept or reject examples generated in response to a query image, is used to locally estimate the strength of features along each dimension while taking into consideration the correlation between features. This results in local neighborhoods that are constricted along feature dimensions that are most relevant, while enlongated along less relevant ones. In addition to exploring and exploiting local principal information, the system seeks a global space for efficient independent feature analysis by combining such local information. We provide experimental results that demonstrate the efficacy of our technique using real-world data.

UR - http://www.scopus.com/inward/record.url?scp=84956976886&partnerID=8YFLogxK

M3 - Conference contribution

SN - 3540665994

SN - 9783540665991

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 103

EP - 115

BT - Machine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings

A2 - Petrou, Maria

A2 - Perner, Petra

PB - Springer Verlag

ER -

Peng J, Bhanu B. Independent feature analysis for image retrieval. In Petrou M, Perner P, editors, Machine Learning and Data Mining in Pattern Recognition - 1st International Workshop, MLDM 1999, Proceedings. Springer Verlag. 1999. p. 103-115. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).