1-9 A Segmentation Method for Texture Images based on Markov Random Field Modeling in Wavelet Feature Space

Abstract

We propose a new segmentation method for Texture images. Our method extract feature of a texture image by wavelet transformation using wavelet frame, and model the feature based on Markov random field model. By the modeling, we can grasp the feature accurately and concisely. Therefore, we expect to be able to segment the texture image accurately.