SPECTRAL COVARIANCE AND FUZZY REGIONS FOR IMAGE INDEXING

Authors
STRICKER M DIMAI A
Citation
M. Stricker et A. Dimai, SPECTRAL COVARIANCE AND FUZZY REGIONS FOR IMAGE INDEXING, Machine vision and applications, 10(2), 1997, pp. 66-73
Citations number
16
Categorie Soggetti
Controlo Theory & Cybernetics","Computer Sciences, Special Topics","Computer Sciences","Engineering, Eletrical & Electronic","Computer Science Cybernetics
Journal title
Machine vision and applications → ACNP
ISSN journal
09328092
Volume
10
Issue
2
Year of publication
1997
Pages
66 - 73
Database
ISI
SICI code
0932-8092(1997)10:2<66:SCAFRF>2.0.ZU;2-B
Abstract
To improve the discrimination power of color-indexing techniques, we e ncode a minimal amount of spatial information in the index, We tessela te each image with five partially overlapping, fuzzy regions, In the i ndex. for each region in an image, we store its average color and the covariance matrix of the color distribution. A similiarity function of these color features is used to match query images with images in the database. In addition, we propose two measures to evaluate the perfor mance of image-indexing techniques. We present experimental results us ing an image database which contains more than 11,600 color images.