A novel level set method based on distribution
Ren Ming-yi, Xiaofeng Li, LI Zai-ming · International Conference on Computational Problem-Solving · 2010
A novel level set method for image segmentation is proposed in this paper. Considering the gray level distribution of pixels in image, a new energy item according to the error rate is introduced. In combination with the penalty item that prevents the level set function from deviating from the signed distance function, a new energy function is obtained. Then the level set evolution equation is deduced by minimize the energy function. The proposed algorithm can deal with the images with blurry boundaries or strong noise. The simulation results show that the proposed method is faster and more effective than traditional level set methods.