Image segmentation based on minimum description length
Wen Furong, Yuan Baozong, Tang Xiaofang · 2005
In this paper, we propose an energy function for image segmentation based on minimum description length. This functional can describe properties of the image as a random field, the region homogeneity properties and the regularity properties of the region boundary. Based on this energy function, we propose the two-step approach to solve this minimization energy function by employing curve evolution, maximum likelihood estimation and level set method.