Kullback-Leibler distance and graph cuts based active contour model for local segmentation

Abstract

Graph cuts based active contour model can obtain the global optimal solution and run faster. To achieve robust local segmentation of inhomogeneous images and improve operation efficiency, a modified Kullback-Leibler distance and graph cuts based active contour (KLD-GCBAC) model in the iterative narrow band frame is proposed in this paper. The dynamic narrow band is constructed between the morphological erosion of evolving curve and the preset outer boundary. In the boundary term of discrete energy formulation, an adaptive coefficient is designed to adjust the weights of edges connecting adjacent pixels. And the Kullback-Leibler distance between the inner and outer local neighborhoods is introduced to the region term as the weight of n-links connecting pixels with terminal nodes. Experimental results on synthetic and medical images with intensity inhomogeneity show that the proposed model can make full use of local features of the image and achieve local segmentation more efficiently.