Non-rigid image registration is an important and challenge work in image processing. The demons algorithm is one of the most effective non-rigid image registration methods. However, it is only suitable for images with small deformation. In recent years, many improving techniques are proposed. The free form deformation method based on B-spline function is widely employed in non-rigid image registration and is good at dealing with large deformation image registration. However, the performance of the demons algorithm is better than that of the B-spline method in dealing with small deformation registration. Therefore, in this paper, we propose to combine the demons algorithm and the B-spline method. The new method consists of two steps: First, it applies the B-spline method to deal with the large deformation. Then, it uses the demons algorithm to treat the small deformation. The testing results show that the new method is effective in dealing images with both small and large deformations. Comparing to the demons algorithm as well as the B-spline method, the new method has the smallest registration error and hence, is the best.
CLAIRE: A Distributed-Memory Solver for Constrained Large Deformation Diffeomorphic Image Registration
