B-spline snakes in two stages

Abstract

In using Snake algorithms, the slow convergence speed is due to the large number of control points to be selected, as well as difficulties in setting the weighting factors that comprise the internal energies of the curve. Even in using the B-Spline snakes, splines cannot be fitted into the corner of the object completely. In this paper, a novel two-stage method based on B-Spline Snakes is proposed. It is superior both in accuracy and fast convergence speed over previous B-Spline Snakes. The first stage reduces the number of control points using potential function V(x,y) minimization. Hence, it allows the spline to quickly approach the minimum energy state. The second stage is designed to refine the B-Spline snakes based on the node points of the polynomials without knots. In other words, an elasticity spline is controlled by node points where knots are fixed. Simulation and validation of results are presented. Compared to the traditional B-Spline snakes, better performance was achieved using the method proposed in this paper.