Supervised texture segmentation using DWT and a modified K-NN classifier

Texture segmentation has been an important problem in image processing. Filtering approaches have been popular, and recent studies have indicated a need for efficient, low-complexity algorithms. In this paper, we present a texture segmentation scheme based on the Discrete Wavelet Transform (DWT). The DWT is a non-redundant representation which can reduce computational complexity in the processing. The texture segmentation scheme presented here consists of three steps: feature extraction, conditioning, and clustering. For feature conditioning, a number of smoothing windows have been tested. Clustering is performed with a modified K-NN clustering algorithm. The proposed scheme consistently achieves error rates of less than 10% with the best average error of 5.62%.