A Novel Salient Image Edge Detection Algorithm Based on Ant Colony Optimization

In this paper, a novel salient image edge detection technique that is based on Ant Colony Optimization (ACO) is presented. Firstly, the proposed method designs a new edge saliency description called Support Region Area (SRA) using phase grouping algorithm. Then, two kind of heuristic information, SRA and gradient magnitude, are introduced in ACO to guide the ants movement. The quantity of pheromone laid by each ant on its new arrived node is calculated based the SRA and the gradient magnitude on the node. Each ants transition probability is calculated by a new method which linear weighted combines the pheromone, the gradient magnitude and the SRA in the ants 8-connectivity neighborhood. A taboo table is created for each ant that recorder the nodes it has recently visited, and is used to present the ant form visiting the same set of nodes repeatedly. Experimental results show the success of the technique in extracting salient edges from visual and infrared images.