Modified Local Updates of the Ant Colony Optimization Algorithm for Image Edge Detection

Abstract

Edge detection refers to the process of extracting edge information of an image. It is considered as a basic step used in the majority of image processing applications. The aim of this study was to modify local updates of pheromones. Therefore, the convergence of the Ant Colony Optimization (ACO) algorithm applied to image edge detection could be accelerated effectively. Such the algorithm is a metaheuristic method applying the ants as agents with their pheromone updates for an effective and efficient solution of search processes. Five ACO algorithms for edge detection, i.e., ACO, modified ACO, ACO with the Sobel operator, ACO with the Prewitt operator, and ACO with the Isotropic operator were in comparison. Nearly optimal solutions of several image datasets were discovered through examination of the number of ants and iterations. Additionally, calculation results of each image dataset and algorithm were compared. The evidence shows that solutions produced by all algorithms are equally good. For an image dataset with more ants, however, it is found that the modified ACO algorithm has the best solution in terms of time convergence. The study contribution is further next to adding the concept of improving edge detection in the image with the ant colony optimization algorithm. The implementation of the study carried out is to modify local updates which are functionally used for improving the edge detection dealt with by ants taking part in ACO.