B-spline representation of active contours

In this paper, we propose two cubic B-spline representations have a lot of advantages such as compact and local control. One is based on registration by a least squares method. The evolutions of an active contour are restricted by an affine transformation which is calculated by corresponding points between pre-deformed and deformed curves during the evolution process. The other is based on dynamic programming. We applied it to control points. It is an energy minimizing spline controlled by control points. We use dynamic programming as an energy minimization method for both approaches. The proposed algorithms is applied to both synthetic binary images and MRI images.