Analysis of multiscale texture segmentation using wavelet-domain hidden Markov models

Wavelet-domain hidden Markov tree (HMT) models are powerful tools for modeling the statistical properties of wavelet transforms. By characterizing the joint statistics of the wavelet coefficients, HMTs efficiently capture the characteristics of a large class of real-world signals and images. In this paper, we apply this multiscale statistical description to the texture segmentation problem. We also show how the Kullback-Leibler (KL) distance between texture models can provide a simple performance indicator.