Abstract:

The variational level set model has been a fundamental framework of multiphase segmentation of images, which is usually solved by computing steady solutions of evolution equations of level set functions. The computing efficiency is low. For the model of piecewise constant image segmentation using n level set functions for n regions, the optimization problem was transformed to discrete binary value level set functions. Then they were transformed to a convex optimization problem. The solutions of the original problems were obtained from threshold solutions of the convex optimization problem. The Split Bregman algorithm was designed for the proposed problem. Some image segmentation examples were presented to prove that the method proposed in this paper improved computing efficiency compared to the traditional method, and one 3D image segmentation example was presented to prove the effectiveness of this method.

Key words:  multiphase image segmentation, variational method, convex optimization method, Split Bregman algorithm