Model-based nonrigid image registration using scale-invariant features

This paper proposes a novel nonrigid medical image registration algorithm based on the finite element elastic model. The deformation is constrained by the image external force and the landmark information while solving the optimization problem. We obtain the landmark information by means of automatically extracting the corresponding scale-invariant features from the reference image and float image. Meanwhile the gradient field of the image similarity metric is derived for the image external force. The qualitative and quantitative evaluations of our proposed method have been performed on both MR image pairs and mammographic images. The result of experiments shows that the algorithm of this paper improves the alignment between the reference image and float image.