A new level set model for unified medical image segmentation

Level set methods (LSMs) have been extensively investigated for medical image segmentation. However, there are a few inherent drawbacks in common level set formulations using image variation or region competition. For example, edge-based LSMs are susceptible to weak or broken boundaries, while region-based ones are often dominated by suboptimal solutions. By incorporating the functional of fuzzy controlling, we propose a new level set model in this paper to combine the merits of edge-based and region-based LSMs while overcoming their drawbacks. It also provides a convenient framework to integrate prior information or knowledge for medical image segmentation. Its performance has been preliminarily verified for medical images of computed tomography (CT) and magnetic resonance imaging (MRI).