[A pyramidal multiphase level set algorithm for multi-object segmentation of 3-D medical images].
Gang Zheng, Huinan Wang · PubMed · 2008
The Chan and Vese (C-V) model using one level set function can only represent one object and one background by the sign of the level set function and is not able to express multiple sub-objects of the object region in three-dimensional (3-D) medical images. To deal with the problem, the two-dimensional (2-D) C-V model is firstly extended to the 3-D model. Secondly, a key technique,designated as the technique of painting background, ensures that the evolving curve stops only inside the object following the theory of simultaneous brightness contrast. Thirdly, a pyramidal multiphase level set algorithm for segmentation is developed by the 3-D C-V model and the proposed technique. Finally, experimental results show that the proposed algorithm is able to partition a given image into multiple objects (n-1 level sets for n objects). Also, the algorithm is especially effective for the detection of sub-objects with weak boundaries.