Image Segmentation Method by Combining Watersheds and Ant Colony Clustering

Abstract

Aimed at resolving the problems of sensitivity to noise and over-segmentation existing in traditional watershed algorithm, we presents a new image segmentation method - combining watersheds and ant colony clustering(CWAC). Firstly, the image is initially segmented using watershed algorithm. Then, ant colony clustering algorithm is used to merge different regions of homogeneity to gain the final result of segmentation. We use intensity and spatial information from watershed transform to define a new visibility which can get more accuracy and efficient clustering ant colony. Experiments show that CWAC algorithm can successfully solve over-segmentation problem and at the same time it can reduce the computational times of ant colony clustering. So CWAC can segment objective quickly and accurately and it is practicable method for the image segmentation.