Image description with features that summarize

J. J. Corso, Gregory Hager

Research output: Contribution to journalArticle

Abstract

We present a new method for describing images for the purposes of matching and registration. We take the point of view that large, coherent regions in the image provide a concise and stable basis for image description. We develop a new algorithm for feature detection that operates on several projections (feature spaces) of the image using kernel-based optimization techniques to locate local extrema of a continuous scale-space of image regions. Descriptors of these image regions and their relative geometry then form the basis of an image description. The emphasis of the work is on features that summarize image content and are highly robust to viewpoint changes and occlusion yet remain discriminative for matching and registration. We present experimental results of these methods applied to the problem of image retrieval. We find that our method performs comparably to two published techniques: Blobworld and SIFT features. However, compared to these techniques two significant advantages of our method are its (1) stability under large changes in the images and (2) its representational efficiency.

Original languageEnglish (US)
Pages (from-to)446-458
Number of pages13
JournalComputer Vision and Image Understanding
Volume113
Issue number4
DOIs
StatePublished - Apr 2009

Fingerprint

Image retrieval
Geometry

Keywords

  • Feature detector
  • Feature space
  • Image matching
  • Interest point operator
  • Segmentation

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Image description with features that summarize. / Corso, J. J.; Hager, Gregory.

In: Computer Vision and Image Understanding, Vol. 113, No. 4, 04.2009, p. 446-458.

Research output: Contribution to journalArticle

@article{2f4f4176626942ef9676eaa530707641,
title = "Image description with features that summarize",
abstract = "We present a new method for describing images for the purposes of matching and registration. We take the point of view that large, coherent regions in the image provide a concise and stable basis for image description. We develop a new algorithm for feature detection that operates on several projections (feature spaces) of the image using kernel-based optimization techniques to locate local extrema of a continuous scale-space of image regions. Descriptors of these image regions and their relative geometry then form the basis of an image description. The emphasis of the work is on features that summarize image content and are highly robust to viewpoint changes and occlusion yet remain discriminative for matching and registration. We present experimental results of these methods applied to the problem of image retrieval. We find that our method performs comparably to two published techniques: Blobworld and SIFT features. However, compared to these techniques two significant advantages of our method are its (1) stability under large changes in the images and (2) its representational efficiency.",
keywords = "Feature detector, Feature space, Image matching, Interest point operator, Segmentation",
author = "Corso, {J. J.} and Gregory Hager",
year = "2009",
month = "4",
doi = "10.1016/j.cviu.2008.11.009",
language = "English (US)",
volume = "113",
pages = "446--458",
journal = "Computer Vision and Image Understanding",
issn = "1077-3142",
publisher = "Academic Press Inc.",
number = "4",

}

TY - JOUR

T1 - Image description with features that summarize

AU - Corso, J. J.

AU - Hager, Gregory

PY - 2009/4

Y1 - 2009/4

N2 - We present a new method for describing images for the purposes of matching and registration. We take the point of view that large, coherent regions in the image provide a concise and stable basis for image description. We develop a new algorithm for feature detection that operates on several projections (feature spaces) of the image using kernel-based optimization techniques to locate local extrema of a continuous scale-space of image regions. Descriptors of these image regions and their relative geometry then form the basis of an image description. The emphasis of the work is on features that summarize image content and are highly robust to viewpoint changes and occlusion yet remain discriminative for matching and registration. We present experimental results of these methods applied to the problem of image retrieval. We find that our method performs comparably to two published techniques: Blobworld and SIFT features. However, compared to these techniques two significant advantages of our method are its (1) stability under large changes in the images and (2) its representational efficiency.

AB - We present a new method for describing images for the purposes of matching and registration. We take the point of view that large, coherent regions in the image provide a concise and stable basis for image description. We develop a new algorithm for feature detection that operates on several projections (feature spaces) of the image using kernel-based optimization techniques to locate local extrema of a continuous scale-space of image regions. Descriptors of these image regions and their relative geometry then form the basis of an image description. The emphasis of the work is on features that summarize image content and are highly robust to viewpoint changes and occlusion yet remain discriminative for matching and registration. We present experimental results of these methods applied to the problem of image retrieval. We find that our method performs comparably to two published techniques: Blobworld and SIFT features. However, compared to these techniques two significant advantages of our method are its (1) stability under large changes in the images and (2) its representational efficiency.

KW - Feature detector

KW - Feature space

KW - Image matching

KW - Interest point operator

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=61349141838&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=61349141838&partnerID=8YFLogxK

U2 - 10.1016/j.cviu.2008.11.009

DO - 10.1016/j.cviu.2008.11.009

M3 - Article

AN - SCOPUS:61349141838

VL - 113

SP - 446

EP - 458

JO - Computer Vision and Image Understanding

JF - Computer Vision and Image Understanding

SN - 1077-3142

IS - 4

ER -