Active contours driven by time-varying fitting energy

Abstract

In this paper, we propose a novel region-based active contour model for images with intensity inhomogeneity. We define an energy functional with both local and global intensity fitting term. A time-varying function is introduced to balance the local and the global intensity term, which combines the advantages of these two terms naturally. Based on this time-varying function, the proposed model possesses both local separability and global consistency. Specifically, local separability helps to tackle the intensity inhomogeneous problem, while global consistency makes the segmentation result insensitive to the initialization. The proposed model is finally incorporated into a level set formulation. Experimental results on both synthetic and real images demonstrate the superior performance of our model.