Cartoon-texture evolution for two-region image segmentation

Abstract Two-region image segmentation is the process of dividing an into two regions interest, i.e., foreground and background. To this aim, Chan et al. (SIAM J Appl Math 66(5):1632–1648, 2006) designed a model well suited for smooth images. One drawback that it may produce bad when contains oscillatory components. Based on cartoon-texture decomposition to be segmented, we propose new able accurate images also containing noise or information like texture. The novel leads non-smooth constrained optimization problem which solve by means ADMM method. convergence numerical scheme proved. Several experiments smooth, noisy, textural show effectiveness proposed model.