A multi-resolution approach for edge detection using Ant Colony Optimization
Abubakar Muhammad Ashir, Alaa Eleyan · 2015
In this paper we aim at providing a robust and more compact approach for detecting edges compared to the traditional edge detection algorithms like Roberts, Sobel, Prewitt and evolutionary-inspired Ant Colony Optimization (ACO) techniques. In this proposed approach, an ACO is used alongside Dual-Tree Complex Wavelet Transform (DT-CWT) to detect and emphasize edges that would have been difficult to obtain with directly applying ACO or conventional edge detection algorithms. Initially the image is decomposed using DT-CWT to obtain the oriented wavelets and approximation versions of the original image. ACO is applied to each of the decomposed images and then image is reconstructed to get the processed image with the detected edges. The results obtained reveal superior, more detailed and emphasized edges than directly applying ACO or other conventional techniques. The proposed approach is also capable of identifying edges in slightly varying intensity regions.