Fast Texture Segmentation Based on Semi-local Region Descriptor and Active Contour Driven by the Bhattacharyya Distance

Abstract

Based on a texture descriptor which intrinsically defines the geometry of textures using semi-local image information and tools from differential geometry, a fast active contour segmentation model for color texture image is proposed. In this model, we use the popular Bhattacharyya distance between the probability density function (pdf) to design the data fitting term which distinguishes the background and textures of interest. Then, a fast algorithm based on the Split-Bregman method is introduced to extract meaningful objects. Finally, some examples on some challenging images are illustrated to verify the possibility of the proposed model.