Multiscale Combination of Physically-Based Registration and Deformation Modeling

1 In this paper we present a novel multiscale approach to recovery of nonrigid motion from sequences of registered intensity and range images. The main idea of our approach is that a nite element (FEM) model can naturally handle both registration and deformation modeling using a single model-driving strategy. The method includes a multiscale iterative algorithm based on analysis of the undirected Hausdor distance to recover correspondences. The method is evaluated with resp ect to speed, accur acy, and noise sensitivity. A dvantages of the proposed approach ar e demonstr ated using man-made elastic materials and human skin motion. Experiments with regular grid featur esare used for performance comparison with a conventional approach (separate snakes and FEM models). It is shown that the new method does not requir ea grid and can adapt the model to available object featur es.