Memetic Image Segmentation Method Based on Digraph Coding

In this paper, we propose a method of digraph coding for memetic image segmentation, which adopts an individual coding based on digraph. Moreover, for image segmentation problems, we design two individual learning strategies which classify the similar vertexes into a class by cutting off the individual from the vertex with the largest weight in the digraph code. In order to evaluate the performance of the new algorithm, we take the segmentation experiment on texture images and remote sensing images and comparing with K-means and FCM. It is proved that our method achieve the image segmentation perfectly by the experiments.