An effective image segmentation approach based on cellular automata principle

In this paper, an effective segmentation approach based on Cellular Automata (CA) principle is proposed for handling images with blurry boundary, complicate structure and high speckle noise. In its energy function, both global image information comparison function and local image information comparison function are developed by extracting different image features. For the evolution environment, two neighborhood systems (von Neumann neighborhood system and Moore neighborhood system) are integrated, and a similarity-based criterion is used for suppressing noise and reducing computational complexity. To evaluate the performance of the proposed method, two kinds of medical images are utilized. The experimental results demonstrate that the proposed method can handle blurry boundaries well, is insensitive to initial condition, robust to noise and segment the BUS images accurately.