Fractional Discrimination for Texture Image Segmentation

Abstract

Texture image segmentation plays an important role in texture analysis. This paper presents an approach to image segmentation by texture classification based on fractional discrimination functions. The idea behind this method is to enhance the texture edge points by means of image decomposition and contextual filtering in terms of the proposed fractional function. In addition, such function is described in a unified form with three-parameters. The parameters determine the global scale in conjunction with local scales for feature identification. Our experimental results show that texture features can be effectively extracted on the basis of the selective fractional discrimination function.