A Hybrid Level Set Segmentation for Medical Imagery
S. Kim, H. Lim · 2006
This article is concerned with a level set segmentation (active contour) algorithm for medical imagery. Due to difficulties such as noise and unclear edges, it is often challenging to obtain a reliable segmentation for medical images. In addition to introducing a new hybrid model which combines a gradient-based model and the Mumford-Shah (gradient-free) method, we study the so-called method of background subtraction (MBS) in order to improve reliability of the new model. A linearized alternating direction implicit method is applied for an efficient time integration. For a fast convergence, we also suggest effective initialization strategies for the level set function. The resulting algorithm has proved to locate the desired edges in 2-4 iterations.