Visual information retrieval system via content-based approach

Hun Woo Yoo, Dong Sik Jang, Seh Hwan Jung, Jin Hyung Park, Kwang Seop Song

Research output: Contribution to journalArticle

46 Citations (Scopus)

Abstract

In this paper a content-based image retrieval method that can search large image databases efficiently by color, texture, and shape content is proposed. Quantized RGB histograms and the dominant triple (hue, saturation, and value), which are extracted from quantized HSV joint histogram in the local image region, are used for representing global/local color information in the image. Entropy and maximum entry from co-occurrence matrices are used for texture information and edge angle histogram is used for representing shape information. Relevance feedback approach, which has coupled proposed features, is used for obtaining better retrieval accuracy. A new indexing method that supports fast retrieval in large image databases is also presented. Tree structures constructed by k-means algorithm, along with the idea of triangle inequality, eliminate candidate images for similarity calculation between query image and each database image. We find that the proposed method reduces calculation up to average 92.2 percent of the images from direct comparison.

Original languageEnglish
Pages (from-to)749-769
Number of pages21
JournalPattern Recognition
Volume35
Issue number3
DOIs
Publication statusPublished - 2002 Mar 1

Fingerprint

Information retrieval systems
Textures
Color
Image retrieval
Entropy
Feedback

Keywords

  • Color
  • Content-based
  • Features
  • Image retrieval
  • Relevance feedback
  • Shape
  • Texture
  • Tree
  • Triangle inequality

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Visual information retrieval system via content-based approach. / Yoo, Hun Woo; Jang, Dong Sik; Jung, Seh Hwan; Park, Jin Hyung; Song, Kwang Seop.

In: Pattern Recognition, Vol. 35, No. 3, 01.03.2002, p. 749-769.

Research output: Contribution to journalArticle

Yoo, Hun Woo ; Jang, Dong Sik ; Jung, Seh Hwan ; Park, Jin Hyung ; Song, Kwang Seop. / Visual information retrieval system via content-based approach. In: Pattern Recognition. 2002 ; Vol. 35, No. 3. pp. 749-769.
@article{46b26fb67868488a99efb743d4a7e57e,
title = "Visual information retrieval system via content-based approach",
abstract = "In this paper a content-based image retrieval method that can search large image databases efficiently by color, texture, and shape content is proposed. Quantized RGB histograms and the dominant triple (hue, saturation, and value), which are extracted from quantized HSV joint histogram in the local image region, are used for representing global/local color information in the image. Entropy and maximum entry from co-occurrence matrices are used for texture information and edge angle histogram is used for representing shape information. Relevance feedback approach, which has coupled proposed features, is used for obtaining better retrieval accuracy. A new indexing method that supports fast retrieval in large image databases is also presented. Tree structures constructed by k-means algorithm, along with the idea of triangle inequality, eliminate candidate images for similarity calculation between query image and each database image. We find that the proposed method reduces calculation up to average 92.2 percent of the images from direct comparison.",
keywords = "Color, Content-based, Features, Image retrieval, Relevance feedback, Shape, Texture, Tree, Triangle inequality",
author = "Yoo, {Hun Woo} and Jang, {Dong Sik} and Jung, {Seh Hwan} and Park, {Jin Hyung} and Song, {Kwang Seop}",
year = "2002",
month = "3",
day = "1",
doi = "10.1016/S0031-3203(01)00072-3",
language = "English",
volume = "35",
pages = "749--769",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "3",

}

TY - JOUR

T1 - Visual information retrieval system via content-based approach

AU - Yoo, Hun Woo

AU - Jang, Dong Sik

AU - Jung, Seh Hwan

AU - Park, Jin Hyung

AU - Song, Kwang Seop

PY - 2002/3/1

Y1 - 2002/3/1

N2 - In this paper a content-based image retrieval method that can search large image databases efficiently by color, texture, and shape content is proposed. Quantized RGB histograms and the dominant triple (hue, saturation, and value), which are extracted from quantized HSV joint histogram in the local image region, are used for representing global/local color information in the image. Entropy and maximum entry from co-occurrence matrices are used for texture information and edge angle histogram is used for representing shape information. Relevance feedback approach, which has coupled proposed features, is used for obtaining better retrieval accuracy. A new indexing method that supports fast retrieval in large image databases is also presented. Tree structures constructed by k-means algorithm, along with the idea of triangle inequality, eliminate candidate images for similarity calculation between query image and each database image. We find that the proposed method reduces calculation up to average 92.2 percent of the images from direct comparison.

AB - In this paper a content-based image retrieval method that can search large image databases efficiently by color, texture, and shape content is proposed. Quantized RGB histograms and the dominant triple (hue, saturation, and value), which are extracted from quantized HSV joint histogram in the local image region, are used for representing global/local color information in the image. Entropy and maximum entry from co-occurrence matrices are used for texture information and edge angle histogram is used for representing shape information. Relevance feedback approach, which has coupled proposed features, is used for obtaining better retrieval accuracy. A new indexing method that supports fast retrieval in large image databases is also presented. Tree structures constructed by k-means algorithm, along with the idea of triangle inequality, eliminate candidate images for similarity calculation between query image and each database image. We find that the proposed method reduces calculation up to average 92.2 percent of the images from direct comparison.

KW - Color

KW - Content-based

KW - Features

KW - Image retrieval

KW - Relevance feedback

KW - Shape

KW - Texture

KW - Tree

KW - Triangle inequality

UR - http://www.scopus.com/inward/record.url?scp=0035294771&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0035294771&partnerID=8YFLogxK

U2 - 10.1016/S0031-3203(01)00072-3

DO - 10.1016/S0031-3203(01)00072-3

M3 - Article

AN - SCOPUS:0035294771

VL - 35

SP - 749

EP - 769

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 3

ER -