16th Conference on Applied Mathematics, Univ. of Central Oklahoma,
Electron. J. Diff. Eqns., Conf. 07, 2001, pp. 61-70.

Determination of the number of texture segments using wavelets

Joseph P. Havlicek & Peter C. Tay

Abstract:
This paper presents a robust method of determining the number of texture segments in an image. We take an $N \times N$ image and decompose it into $N \times N$ blocks. A three-scale two-dimensional discrete wavelet transform is performed on each $N \times N$ block. For each $N \times N$ block, this transformation produces coefficients for 25 wavelet channels. The energy of each channel is used as a tuple for a vector in the feature space. Nearest neighbor clustering is used to segment the feature space. A measure is defined to determine the "goodness" of the clustering. The optimal number of segments is taken to be the clustering which maximizes our measure.

Published July 20, 2001.
Subject lassfications: 65T60, 42C40.
Key words: Differential geometry, algebraic geometry.

Show me the PDF file (517K), TEX file, and other files for this article.

Joseph P. Havlicek
School of Electrical & Computer Engineering
The University of Oklahoma
Norman, OK 73019-1023 USA
e-mail: joebob@ou.edu
Peter C. Tay
School of Electrical & Computer Engineering
The University of Oklahoma
Norman, OK 73019-1023 USA
e-mail: Peter.C.Tay-1@ou.edu

Return to the EJDE web page