Joint estimation of cardiac kinematics and material parameters from noisy imaging data and uncertain mechanical model

There have been many efforts using image analysis algorithms to study cardiac kinematics, or using biomechanics strategies to study myocardial material properties. In this paper, we propose a novel stochastic mechanics strategy and an extended Kalman filter (EKF) computational framework to estimate the cardiac kinematics functions and material model parameters simultaneously, given a particular a priori myocardial material model with uncertain parameters and a posteriori noisy imaging/imaging-derived data. We address the issues concerning the data-dependent uncertainty of the constraining mechanical models (and their parameters), which are needed in the ill-posed problems. Because of the periodic nature of the cardiac dynamics, we conclude experimentally that it is possible to adopt this physical-model based optimal estimation approach to achieve converged estimates. Results from canine MR phase contrast images and linear elastic model are presented.