Expert system based on artificial neural networks for content-based image retrieval

Sang Sung Park, Kwang K. Seo, Dong Sik Jang

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

Clustering technique is essential for fast retrieval in large database. In this paper, new image clustering technique based on artificial neural networks is proposed for content-based image retrieval. Fuzzy-ART mechanism maps high-dimensional input features into the output neuron. Joint HSV histogram and average entropy computed from gray-level co-occurrence matrices in the localized image region is employed as input feature elements. Original Fuzzy-ART suffers unnecessary increase of the number of output neurons when the noise input is presented. Modified Fuzzy-ART mechanism resolves the problem by differently updating the committed node and uncommitted node, and checking the vigilance test again. To show the validity of the proposed algorithm, experiment results on image clustering performance and comparison with original Fuzzy-ART are presented in terms of recall rates.

Original languageEnglish
Pages (from-to)589-597
Number of pages9
JournalExpert Systems with Applications
Volume29
Issue number3
DOIs
Publication statusPublished - 2005 Oct 1

Fingerprint

Image retrieval
Expert systems
Neurons
Neural networks
Entropy
Experiments

Keywords

  • Content-based image retrieval
  • Fuzzy-ART
  • Gray-level co-occurrence matrix
  • HSV joint histogram
  • Image clustering

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications

Cite this

Expert system based on artificial neural networks for content-based image retrieval. / Park, Sang Sung; Seo, Kwang K.; Jang, Dong Sik.

In: Expert Systems with Applications, Vol. 29, No. 3, 01.10.2005, p. 589-597.

Research output: Contribution to journalArticle

@article{14086918eb554231b04fcd69333e274f,
title = "Expert system based on artificial neural networks for content-based image retrieval",
abstract = "Clustering technique is essential for fast retrieval in large database. In this paper, new image clustering technique based on artificial neural networks is proposed for content-based image retrieval. Fuzzy-ART mechanism maps high-dimensional input features into the output neuron. Joint HSV histogram and average entropy computed from gray-level co-occurrence matrices in the localized image region is employed as input feature elements. Original Fuzzy-ART suffers unnecessary increase of the number of output neurons when the noise input is presented. Modified Fuzzy-ART mechanism resolves the problem by differently updating the committed node and uncommitted node, and checking the vigilance test again. To show the validity of the proposed algorithm, experiment results on image clustering performance and comparison with original Fuzzy-ART are presented in terms of recall rates.",
keywords = "Content-based image retrieval, Fuzzy-ART, Gray-level co-occurrence matrix, HSV joint histogram, Image clustering",
author = "Park, {Sang Sung} and Seo, {Kwang K.} and Jang, {Dong Sik}",
year = "2005",
month = "10",
day = "1",
doi = "10.1016/j.eswa.2005.04.027",
language = "English",
volume = "29",
pages = "589--597",
journal = "Expert Systems with Applications",
issn = "0957-4174",
publisher = "Elsevier Limited",
number = "3",

}

TY - JOUR

T1 - Expert system based on artificial neural networks for content-based image retrieval

AU - Park, Sang Sung

AU - Seo, Kwang K.

AU - Jang, Dong Sik

PY - 2005/10/1

Y1 - 2005/10/1

N2 - Clustering technique is essential for fast retrieval in large database. In this paper, new image clustering technique based on artificial neural networks is proposed for content-based image retrieval. Fuzzy-ART mechanism maps high-dimensional input features into the output neuron. Joint HSV histogram and average entropy computed from gray-level co-occurrence matrices in the localized image region is employed as input feature elements. Original Fuzzy-ART suffers unnecessary increase of the number of output neurons when the noise input is presented. Modified Fuzzy-ART mechanism resolves the problem by differently updating the committed node and uncommitted node, and checking the vigilance test again. To show the validity of the proposed algorithm, experiment results on image clustering performance and comparison with original Fuzzy-ART are presented in terms of recall rates.

AB - Clustering technique is essential for fast retrieval in large database. In this paper, new image clustering technique based on artificial neural networks is proposed for content-based image retrieval. Fuzzy-ART mechanism maps high-dimensional input features into the output neuron. Joint HSV histogram and average entropy computed from gray-level co-occurrence matrices in the localized image region is employed as input feature elements. Original Fuzzy-ART suffers unnecessary increase of the number of output neurons when the noise input is presented. Modified Fuzzy-ART mechanism resolves the problem by differently updating the committed node and uncommitted node, and checking the vigilance test again. To show the validity of the proposed algorithm, experiment results on image clustering performance and comparison with original Fuzzy-ART are presented in terms of recall rates.

KW - Content-based image retrieval

KW - Fuzzy-ART

KW - Gray-level co-occurrence matrix

KW - HSV joint histogram

KW - Image clustering

UR - http://www.scopus.com/inward/record.url?scp=24144470789&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=24144470789&partnerID=8YFLogxK

U2 - 10.1016/j.eswa.2005.04.027

DO - 10.1016/j.eswa.2005.04.027

M3 - Article

AN - SCOPUS:24144470789

VL - 29

SP - 589

EP - 597

JO - Expert Systems with Applications

JF - Expert Systems with Applications

SN - 0957-4174

IS - 3

ER -