Regularization of phase contrast magnetic resonance images using optical flow and smoothness constraints

This paper presents a post processing strategy for myocardial velocity fields obtained by phase contrast magnetic resonance imaging. Such data can be used to track cardiac motion and to calculate strain. The method combines data regularization with optical flow estimation to overcome the partial volume effect in the image acquisition. Validation is performed both in vitro and in vivo and it is shown that the method improves the accuracy of cardiac motion tracking