Ant colony optimization for the K-means algorithm in image segmentation

Abstract

In this paper the ant colony optimization (ACO) is used in the K-means algorithm for improving the image segmentation. The learning mechanism of this algorithm is formulated by using the ACO meta-heuristic. As the pheromone dominates the exploration of ants for problem solutions, preliminary experiments on pheromone's update are reported. Two methods for defining and updating pheromone values are proposed and tested: one with the spatial coordinate distances and the other without using such a distance. The ACO improves the K-means algorithm by making it less dependent on the initial parameters.