Variational image segmentation on implicit surface using Split-Bregman method

The coupling images and their underlying surfaces results in complex implementation and low computing efficiency of image segmentation on surfaces. For the piecewise constant and smooth image segmentation on surface, the traditional Chan-Vese models are transformed to variational level set models on implicit surfaces and computed by using fast Split-Bregman methods in this paper. Additionally, the Split-Bregman methods are implemented based on the corresponding globally convex models to avoid the effects of contour initialization in segmentation results. Comparisons of experiment results validate the superiority of the models and algorithms presented in this paper.