Learning landmark geodesics using the ensemble Kalman filter

<p style='text-indent:20px;'>We study the problem of diffeomorphometric geodesic landmark matching where objective is to find a diffeomorphism that, via its group action, maps between two sets landmarks. It well-known that motion landmarks, and thereby diffeomorphism, can be encoded by an initial momentum leading formulation solved as optimisation over such momenta. The novelty our work lies in application derivative-free Bayesian inverse method for learning optimal encoding diffeomorphic mapping template target. we apply ensemble Kalman filter, extension filter nonlinear operators. We describe efficient implementation algorithm show several numerical results various target shapes.</p>