Digital Image Edge Detection Using Directional Ant Colony Optimization Based on Gradient Magnitude and Direction

Ant Colony Optimization (ACO) is a method that imitates the foraging behavior of ants that can be applied to improve the edge detection. Generally, pheromone of ants is guided by local variation in image intensity which is less sensitive for detect edge, thus we need addition of edge information. In this study we propose Directional ACO (DACO) which uses the addition of edge information based on gradient magnitude and direction. In the proposed method, the weight of gradient magnitude and directional initialized firstly, and then ant construct edge using probabilistic which is not only considered by pheromone and local variation of intensity, but also gradient magnitude and direction. in the each iteration, the edge is constructed by applying a threshold using Otsu. Final edge is determined if the difference of edge number has reach a threshold. Experiments were conducted using images from private synthetic dataset and CID’s natural image dataset. Figure of merit was used to evaluate quantitatively performance of the proposed method. The experiment showed that DACO reached 0.812 (81.2%), whereas standard ACO reached 0.494 (49.4%). Experiment results showed that DACO outperforms standard ACO.