Nonlinear registration of brain images using deformable models

A key issue in several brain imaging applications, including computer aided neurosurgery, functional image analysis, and morphometrics, is the spatial normalization and registration of tomographic images from different subjects. This paper proposes a technique for spatial normalization of brain images based on elastically deformable models. In the authors' approach they use a deformable surface algorithm to find a parametric representation of the outer cortical surface and then use this representation to obtain a map between corresponding regions of the outer cortex in two different images. Based on the resulting map, the authors then derive a three-dimensional elastic warping transformation which brings two images in register. This transformation models images as inhomogeneous elastic objects which are deformed into registration with each other by external force fields. The elastic properties of the images can vary from one region to the other, allowing more variable brain regions, such as the ventricles, to deform more freely than less variable ones. Finally, the authors use prestrained elasticity to model structural irregularities, and in particular the ventricular expansion occurring with aging or diseases. The performance of the authors' algorithm is demonstrated on magnetic resonance images.