Natural image segmentation based on contour detection
Meng Yang · 2012
In this paper,we propose a new method of natural image segmentation based on contour detection.The algorithms combine the statistical techniques of minimizing MDL(Minimum Description Length) criterion which involves the whole image and the active contour model in a variational level set formulation.This method is effective to segment the most salient objects in the scene.Comparisons with other state-of-the-art approaches validate our method's advantages.