Inverse Consistent Deformable Image Registration

This paper presents a novel variational model for inverse consistent deformable image registration. The proposed model deforms both source and target images simultaneously, and aligns the deformed images in the way that the forward and backward transformations are inverse consistent. To avoid the direct computation of the inverse transformation fields, our model estimates two more vector fields by minimizing their invertibility error using the deformation fields. Moreover, to improve the robustness of the model to the choice of parameters, the dissimilarity measure in the energy functional is derived using the likelihood estimation. The experimental results on clinical data indicate the efficiency of the proposed method with improved robustness, accuracy and inverse consistency.