Finite difference based algorithms for optimal snakes used in medical diagnosis

In this paper, we describe briefly the 2D parametric deformable models, then we infer an algorithm, via Crank-Nicolson finite difference scheme for finding the so-called energy-minimizing (i.e. optimal) snake. The residue, the truncature - error and the order of accuracy related to this algorithm are estimated, in order to establish its consistency. We prove that the algorithm leads to better results regarding its convergence than the algorithm developed by the explicit finite difference scheme. Some aspects concerning the computerized assistance, that implement this algorithm in both prosthetic surgical methods and medical treatment monitoring in Metabolic Syndrome, are also pointed out.