Edge detection by evolutionary optimization and evolutionary minimization of cost function

In this paper we present an edge detection method based on evolutionary techniques. The edge detection is casted as an optimization problem which minimizes cost function. Similar to the Bhandarkar's approach, the cost function is constrained by accurate localization, thinness, curvature, continuity and length. However, the proposed method uses evolutionary techniques not only for the search of minimum of the cost function, but also for the formulation of optimal structure of cost function. The usefulness of this method is demonstrated by showing experimental results of detected edges.