Image Segmentation Based On Arificial Immune Network

Artificial Immune System (AIS) was an adaptive system that used to solve complex problem. One of the AIS Artificial Immune Network was used to analysis data clustering problem. Data clustering was a normal method to segment image. So this thesis mainly research on the image segmentation based on Artificial Immune Network.The main works of this thesis are as follow:(1) An image segmentation algorithm based on Multi-Valued Immune Network (MVIN) and FCM was proposed. Because of FCM algorithm was sensitive to initial clustering centers. MVIN was used to train the clustering centers to relieve the problem. Then It could be used to segment image. And this method was effectively applied to Brodatz Texture and SAR image segmentation.(2) An SAR image segmentation algorithm based on watershed algorithm and Artificial Immune Network (AINET) was proposed. Because of the AINET has great complexity in data clustering, it will waste lots of time and resources to segment image using AINET directly. In order to solve this problem, we used watershed algorithm to segment SAR image firstly. The image was made into lots of small blocks, and each block was chosen as a sample. Then we use the AINET to cluster the samples of image to segment image. The feasibility of the algorithm has been proved by the experiments.