Texture segmentation based on incomplete wavelet packet frame

In this paper, we design a new method to do texture segmentation. We adopt the incomplete wavelet packet frame to decompose the texture image into multichannel subimages. These subimages contain the middle frequency information where the main information of texture focuses and describe the feature of every pixel which is very helpful for the pixel based exact segmentation as well. To these subimages different feature extraction methods are applied according to the different character of themselves. Energy and entropy features are extracted from the detail subimages. In order to smooth the fragment caused by the detail feature, we extract the mean variance feature from the approximate subimages. We propose a quadrant mean filter to smooth the feature's noise and alleviate the boundary effect of the feature. The fuzzy c-means is provided to do classification of these features. The performance of this new method is demonstrated on the segmentation of Brodatz textures.