Segmentation of 3D brain structures using level sets and dense registration

Presents a cooperative strategy for the segmentation of 3D brain MRI which integrates 3D segmentation and 3D registration methods. The segmentation is based on the level set formalism. A closed 3D surface representing the structure of interest iteratively propagates towards the desired boundaries through the evolution of a 4D implicit function. In this work, an adaptive propagation direction depending on local intensity values is used. In addition, an adaptive iteration step is automatically computed at each iteration in order to improve the robustness and the efficiency of the algorithm. The main contribution of this work is the use of an automatic registration method to initialize the surface, as an alternative solution to manual initialization. Registration is achieved through a robust multiresolution and multigrid minimization scheme. This cooperation significantly improves the quality of the method, since the segmentation is faster and fully automatic. Results on volumetric brain MR images are presented and discussed.