Region Segmentation via Deformable Model-Guided Split and Merge

摘要

An improved method for deformable shape based image segmentation is described. Image regions are merged together and/or split apart, based on their agreement with an a priori distribution on the global deformation parameters for a shape template. Perceptually motovated crtieria are used to determine where/how to split regions, based on the local shape properties of the regions group's bounding contour. A globally consistent interpretation is determined in part by the minimum description lingth principle. Experiments show that model guided split and merge yields a significant improvement in segmentation over a method that uses merging alone.