Fast non-rigid registration of medical images: comparing the thin-plate and biharmonic splines

The merging of multi-modal medical image data, typically used for applications such as surgical guidance and computer aided diagnosis, has become increasingly important. In particular, the speed by which such data can be registered is often crucial. This paper presents a novel approach based on fast Radial Basis Function evaluation using the thin-plate and biharmonic spline models. The new algorithm is evaluated in software and hardware assisted versions and runs on a standard PC. We present a number of experiments on synthetic data illustrating substantial speedups for both spline models. We also report an initial experiment on the registration of MR and CT data of size 256 which yields sub-second performance on a standard PC.