A comparative study on the optimization of global optimum active contours
Jonas De Vylder, Jan Aelterman, Koen Douterloigne, Wilfried R. Philips · 2011
The active contour framework is widely used for segmentation of images. Specially active contours with a convex energy gained a lot of interest, since they are independent of initialization. Several methods have been proposed to minimise the active contour energy, all claiming to be fast and efficient. In this paper we study and compare the convergence and constraints of these methods. We also propose a new optimization method specially designed for the optimization of convex energy active contours. The proposed method is a combination of a line search method with a Newton-Raphson optimization technique. The proposed method shows good convergence results, while maintaining a low computation cost.