Leaking prevention in fast level sets using fuzzy models: an application in MR brain

One of the weakness of classical deformable models and level sets is the boundary leaking during the curve propagation. This is because of weak stopping or clamping forces in noisy images. This paper presents a level-set method which uses fuzzy models to prevent leaking. The forces in the level-set approach use four kinds of speed control functions based on shape, region, edge and curvature. Regional and shape speed functions were determined based on the fuzzy membership function computed using the fuzzy clustering technique, while the edge and curvature speed functions were based on gradient and signed distance transform functions, respectively. The level-set algorithm was implemented to run in the “narrow band” using a “fast marching method”. The system was tested on synthetic convoluted shapes and real magnetic resonance images of the human head. The entire system took around a minute to estimate the white matter/gray matter boundaries on an XP1000 running the Linux operating system when the raw contour was placed halfway from the goal, and it took a few seconds if the raw contour was placed closed to the goal boundary with 100% accuracy