An elasticity-based region model and its application to the estimation of the heart deformation in tagged MRI

We propose a novel deformable model to assess the deformation of a textured object in an image sequence by simultaneously tracking edge and intensity information. The proposed model is defined as an elastic region with snake-like rigidity constraints at its boundaries. By regularizing the displacement estimation with an elasticity-based constraint, this model is able to assess physically realistic deformations inside a region. By incorporating rigidity constraint at its boundaries, it can also accurately track edges. The numerical implementation of the model is performed using the finite element method. The good behavior of the proposed model is illustrated on synthetic images which simulates the heart contraction in tagged magnetic resonance imaging (MRI). The contour tracking abilities of the model are also illustrated on standard MR images