Learning similarity measure for natural image retrieval with relevance feedback

Guo-Dong Guo; Anil K. Jain; Wei-Ying Ma; Hong-Jiang Zhang

doi:10.1109/TNN.2002.1021882

Learning similarity measure for natural image retrieval with relevance feedback

Guo-Dong Guo
, Anil K. Jain
, Wei-Ying Ma
, Hong-Jiang Zhang

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

199 Citations (Scopus)

Abstract

A new scheme of learning similarity measure is proposed for content-based image retrieval (CBIR). It learns a boundary that separates the images in the database into two clusters. Images inside the boundary are ranked by their Euclidean distances to the query. The scheme is called constrained similarity measure (CSM), which not only takes into consideration the perceptual similarity between images, but also significantly improves the retrieval performance of the Euclidean distance measure. Two techniques, support vector machine (SVM) and AdaBoost from machine learning, are utilized to learn the boundary. They are compared to see their differences in boundary learning. The positive and negative examples used to learn the boundary are provided by the user with relevance feedback. The CSM metric is evaluated in a large database of 10 009 natural images with an accurate ground truth. Experimental results demonstrate the usefulness and effectiveness of the proposed similarity measure for image retrieval.
© 2002 IEEE

Original language	English
Pages (from-to)	811-820
Journal	IEEE Transactions on Neural Networks
Volume	13
Issue number	4
DOIs	https://doi.org/10.1109/TNN.2002.1021882
Publication status	Published - Jul 2002
Externally published	Yes

Bibliographical note

Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].

Research Keywords

AdaBoost
Constrained similarity measure
Content-based image retrieval
Feature selection
Learning
Relevance feedback
Support vector machine (SVM)

Access Output

10.1109/TNN.2002.1021882

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@article{03a27bba927745e8b20c38f7949cba28,

title = "Learning similarity measure for natural image retrieval with relevance feedback",

abstract = "A new scheme of learning similarity measure is proposed for content-based image retrieval (CBIR). It learns a boundary that separates the images in the database into two clusters. Images inside the boundary are ranked by their Euclidean distances to the query. The scheme is called constrained similarity measure (CSM), which not only takes into consideration the perceptual similarity between images, but also significantly improves the retrieval performance of the Euclidean distance measure. Two techniques, support vector machine (SVM) and AdaBoost from machine learning, are utilized to learn the boundary. They are compared to see their differences in boundary learning. The positive and negative examples used to learn the boundary are provided by the user with relevance feedback. The CSM metric is evaluated in a large database of 10 009 natural images with an accurate ground truth. Experimental results demonstrate the usefulness and effectiveness of the proposed similarity measure for image retrieval. {\textcopyright} 2002 IEEE",

keywords = "AdaBoost, Constrained similarity measure, Content-based image retrieval, Feature selection, Learning, Relevance feedback, Support vector machine (SVM)",

author = "Guo-Dong Guo and Jain, \{Anil K.\} and Wei-Ying Ma and Hong-Jiang Zhang",

note = "Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].",

year = "2002",

month = jul,

doi = "10.1109/TNN.2002.1021882",

language = "English",

volume = "13",

pages = "811--820",

journal = "IEEE Transactions on Neural Networks",

issn = "1045-9227",

publisher = "IEEE Computational Intelligence Society",

number = "4",

}

TY - JOUR

T1 - Learning similarity measure for natural image retrieval with relevance feedback

AU - Guo, Guo-Dong

AU - Jain, Anil K.

AU - Ma, Wei-Ying

AU - Zhang, Hong-Jiang

N1 - Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].

PY - 2002/7

Y1 - 2002/7

N2 - A new scheme of learning similarity measure is proposed for content-based image retrieval (CBIR). It learns a boundary that separates the images in the database into two clusters. Images inside the boundary are ranked by their Euclidean distances to the query. The scheme is called constrained similarity measure (CSM), which not only takes into consideration the perceptual similarity between images, but also significantly improves the retrieval performance of the Euclidean distance measure. Two techniques, support vector machine (SVM) and AdaBoost from machine learning, are utilized to learn the boundary. They are compared to see their differences in boundary learning. The positive and negative examples used to learn the boundary are provided by the user with relevance feedback. The CSM metric is evaluated in a large database of 10 009 natural images with an accurate ground truth. Experimental results demonstrate the usefulness and effectiveness of the proposed similarity measure for image retrieval. © 2002 IEEE

AB - A new scheme of learning similarity measure is proposed for content-based image retrieval (CBIR). It learns a boundary that separates the images in the database into two clusters. Images inside the boundary are ranked by their Euclidean distances to the query. The scheme is called constrained similarity measure (CSM), which not only takes into consideration the perceptual similarity between images, but also significantly improves the retrieval performance of the Euclidean distance measure. Two techniques, support vector machine (SVM) and AdaBoost from machine learning, are utilized to learn the boundary. They are compared to see their differences in boundary learning. The positive and negative examples used to learn the boundary are provided by the user with relevance feedback. The CSM metric is evaluated in a large database of 10 009 natural images with an accurate ground truth. Experimental results demonstrate the usefulness and effectiveness of the proposed similarity measure for image retrieval. © 2002 IEEE

KW - AdaBoost

KW - Constrained similarity measure

KW - Content-based image retrieval

KW - Feature selection

KW - Learning

KW - Relevance feedback

KW - Support vector machine (SVM)

UR - https://www.scopus.com/pages/publications/0036650802

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-0036650802&origin=recordpage

U2 - 10.1109/TNN.2002.1021882

DO - 10.1109/TNN.2002.1021882

M3 - RGC 21 - Publication in refereed journal

SN - 1045-9227

VL - 13

SP - 811

EP - 820

JO - IEEE Transactions on Neural Networks

JF - IEEE Transactions on Neural Networks

IS - 4

ER -

Learning similarity measure for natural image retrieval with relevance feedback

Abstract

Bibliographical note

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this