An Adaptive Evolution ACO Algorithm for Image Segmentation

In this paper, an improved algorithm on ant colony optimization with adaptive evolution is proposed for image segmentation. The algorithm combines the deterministic selection strategy and the stochastic one. And this algorithm searches the global optimal solution adaptively adjusting the probability of deterministic selection. After the ants evolved several generations, the pheromone is adjusted by shortening the difference between the best path and the worst one and by properly increasing the probability of stochastic selection. Thus, speed-up is achieved and the problem of stagnation is overcome. The experiments show that the improved algorithm has better performance in image segmentation and the time is largely saved.