In recent years, some variational level set formulations, which are composed of a certain energy functional term and a distance regularization term(DRT), have been proposed. During the evolution of the level set function(LSF), the DRT maintains the LSF stability, eliminating the need for the complex re-initialization procedures of the traditional level set methods. The DRT should have different coefficients in different regions of the LSF. However, there is only one coefficient for the DRT of the existing re-initialization free algorithms. So we propose a flexible coefficient distance regularized level set evolution where the DRT has different coefficients in different regions of the LSF, so that the evolve rate of the LSF and image segmentation performance can be optimized. Experiments show that our algorithm has achieved good results.
YU Ronggui (Thu,) studied this question.