Texture image segmentation using fully global minimization active contour model

Abstract

A new global minimization active contour model is proposed, which has three advantages compared to other active contours. Firstly, the energy function of proposed model is convex, so the proposed model is not sensitive to the initial condition because of having no existence of local minimum in the active contour energy; Secondly by combining the gray levels of pixels and texture information of an image, this method can be used for segmentation of a texture image or a none-texture image. Finally, LBP (local binary pattern) is employed to extract texture features, so computation complexity of proposed model is low. The segmentation tests for synthetic and SAR texture images show that the proposed segmentation model is efficient, accurate, fast and robust.