Dense non-rigid motion estimation in sequences of 3D images using differential constraints

Abstract

We describe a new method for computing the displacement vector field in time sequences of 3D images (4D data). The method is an energy-minimizing method; the energy is splitted into two terms, with one term matching differential singularities in the images, and the other constraining the regularity of the field. In order to reduce the computational time, we introduce an adaptive volume mesh the resolution of which depends on the presence of high gradient. We next perform a modal analysis, which allows a compact representation of the deformation by a reduced number of parameters. We present experimental results on synthetic data and on medical images.