Active contour based on different mathematic statistics

Traditional active contour models using global region information are often based on two constants which can approximate the image gray inside and outside the curves. Those constants can be considered as the one of the mathematic statistics data according to the corresponding region information of the image. In this paper, we propose a different energy which bases on two different types of mathematic statistics. We argue that when those two different types of mathematic statistics are equal, the energy will achieve its minimum. Experiments for synthetic and nature images get desirable results.