Kernel vector approximation files for relevance feedback retrieval in large image databases

Douglas R. Heisterkamp, Jing Peng

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

Many data partitioning index methods perform poorly in high dimensional space and do not support relevance feedback retrieval. The vector approximation file (VA-File) approach overcomes some of the difficulties of high dimensional vector spaces, but cannot be applied to relevance feedback retrieval using kernel distances in the data measurement space. This paper introduces a novel KVA-File (kernel VA-File) that extends VA-File to kernel-based retrieval methods. An efficient approach to approximating vectors in an induced feature space is presented with the corresponding upper and lower distance bounds. Thus an effective indexing method is provided for kernel-based relevance feedback image retrieval methods. Experimental results using large image data sets (approximately 100,000 images with 463 dimensions of measurement) validate the efficacy of our method.

Original languageEnglish
Pages (from-to)175-189
Number of pages15
JournalMultimedia Tools and Applications
Volume26
Issue number2
DOIs
StatePublished - 1 Jun 2005

Fingerprint

Feedback
Image retrieval
Vector spaces

Keywords

  • Content-based image retrieval
  • Indexing
  • Kernel methods
  • Relevance feedback
  • VA-File

Cite this

@article{2fabdc05071946758ff3d14ca33c517b,
title = "Kernel vector approximation files for relevance feedback retrieval in large image databases",
abstract = "Many data partitioning index methods perform poorly in high dimensional space and do not support relevance feedback retrieval. The vector approximation file (VA-File) approach overcomes some of the difficulties of high dimensional vector spaces, but cannot be applied to relevance feedback retrieval using kernel distances in the data measurement space. This paper introduces a novel KVA-File (kernel VA-File) that extends VA-File to kernel-based retrieval methods. An efficient approach to approximating vectors in an induced feature space is presented with the corresponding upper and lower distance bounds. Thus an effective indexing method is provided for kernel-based relevance feedback image retrieval methods. Experimental results using large image data sets (approximately 100,000 images with 463 dimensions of measurement) validate the efficacy of our method.",
keywords = "Content-based image retrieval, Indexing, Kernel methods, Relevance feedback, VA-File",
author = "Heisterkamp, {Douglas R.} and Jing Peng",
year = "2005",
month = "6",
day = "1",
doi = "10.1007/s11042-005-0454-4",
language = "English",
volume = "26",
pages = "175--189",
journal = "Multimedia Tools and Applications",
issn = "1380-7501",
publisher = "Springer Netherlands",
number = "2",

}

Kernel vector approximation files for relevance feedback retrieval in large image databases. / Heisterkamp, Douglas R.; Peng, Jing.

In: Multimedia Tools and Applications, Vol. 26, No. 2, 01.06.2005, p. 175-189.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Kernel vector approximation files for relevance feedback retrieval in large image databases

AU - Heisterkamp, Douglas R.

AU - Peng, Jing

PY - 2005/6/1

Y1 - 2005/6/1

N2 - Many data partitioning index methods perform poorly in high dimensional space and do not support relevance feedback retrieval. The vector approximation file (VA-File) approach overcomes some of the difficulties of high dimensional vector spaces, but cannot be applied to relevance feedback retrieval using kernel distances in the data measurement space. This paper introduces a novel KVA-File (kernel VA-File) that extends VA-File to kernel-based retrieval methods. An efficient approach to approximating vectors in an induced feature space is presented with the corresponding upper and lower distance bounds. Thus an effective indexing method is provided for kernel-based relevance feedback image retrieval methods. Experimental results using large image data sets (approximately 100,000 images with 463 dimensions of measurement) validate the efficacy of our method.

AB - Many data partitioning index methods perform poorly in high dimensional space and do not support relevance feedback retrieval. The vector approximation file (VA-File) approach overcomes some of the difficulties of high dimensional vector spaces, but cannot be applied to relevance feedback retrieval using kernel distances in the data measurement space. This paper introduces a novel KVA-File (kernel VA-File) that extends VA-File to kernel-based retrieval methods. An efficient approach to approximating vectors in an induced feature space is presented with the corresponding upper and lower distance bounds. Thus an effective indexing method is provided for kernel-based relevance feedback image retrieval methods. Experimental results using large image data sets (approximately 100,000 images with 463 dimensions of measurement) validate the efficacy of our method.

KW - Content-based image retrieval

KW - Indexing

KW - Kernel methods

KW - Relevance feedback

KW - VA-File

UR - http://www.scopus.com/inward/record.url?scp=21244480926&partnerID=8YFLogxK

U2 - 10.1007/s11042-005-0454-4

DO - 10.1007/s11042-005-0454-4

M3 - Article

AN - SCOPUS:21244480926

VL - 26

SP - 175

EP - 189

JO - Multimedia Tools and Applications

JF - Multimedia Tools and Applications

SN - 1380-7501

IS - 2

ER -