Canny edge detector improvement using an intelligent ants routing

Abstract

Canny operator provides simple technique to extract the useful edge information from image. However, it suffers from two problems: a difficulty to choose the thresholds values and the presence of broken edges. This paper proposes to improve Canny detector in two steps: First, adaptive Otsu threshold is used to select appropriate thresholds. Then, a new variant of ant colony optimization (ACO) algorithm is used to recover missing edges. From binary image, extracted edge endpoints are used as starting pixels for an intelligent ants routing mechanism. Ants are directed to suitable contour areas and missing edges are retraced via the pheromone traces. Proposed improvements are evaluated with entropy and kappa methods. Experimental results are good and assert approach’s ability to recover most of the broken edges, particularly in noisy clinical images.