Abstract
Abstract. Edges contain important information in image and edge detection can be considered a low level process in image processing. Among different methods developed for this purpose traditional methods are simple and rather efficient. In Swarm Intelligent methods developed in last decade, ACO is more capable in this process. This paper uses traditional edge detection operators such as Sobel and Canny as input to ACO and turns overall process adaptive to application. Magnitude matrix or edge image can be used for initial pheromone and ant distribution. Image size reduction is proposed as an efficient smoothing method. A few parameters such as area and diameter of travelled path by ants are converted into rules in pheromone update process. All rules are normalized and final value is acquired by averaging.
Highlights
Edges show relative sudden changes in image and usually contain important information
This paper proposes an adjusted Ant Colony Optimization (ACO) algorithm in which initial conditions are especially acquired using traditional edge detection techniques
Edges in image contain important information and edge detection plays an important role in image processing
Summary
Edges show relative sudden changes in image and usually contain important information. Ant Colony Optimization (ACO) is a heuristic search method that operates based on ant colony and is being used for discontinuous problems During recent years this method is especially developed for edge extraction purposes (Agarwal, 2012). (Koschan and Abidi, 2005) review edge classification which is a high level process and a connector to further processes such as segmentation Using this classification itself requires hybrid techniques in which edge detection is no longer a pre-processing but part of a hybrid main process to achieve best result using all means. In this case contextual methods which usually combine with Neural Network or fuzzy are applied (Yu and Chang, 2006).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
