RESUMO
In this paper, a novel registration method is presented for three-dimensional mapping of cardiac chambers and computed tomography (CT) segmented surfaces. The proposed method applies random perturbations to three-dimensional mapping of cardiac chambers in order to move out of local minima. It effectively solves the problem of the existing algorithm based on iterative closest point (ICP) in which the optimization often falls into local minima, and improves the registration accuracy and success rate. The registration experiment results show that the final registration accuracy and success rate of the proposed method are better than the existing ICP-based algorithm, and it can implement the registration when the three-dimensional mapping of cardiac chambers and CT segmented surface have a wide range of initial offset.