University of California San Francisco
The authors introduce an algorithm for tracking deformable objects from a sequence of point clouds. The proposed tracking algorithm is based on a probabilistic generative model that incorporates observations of the point cloud and the physical properties of the tracked object and its environment. They propose a modified expectation maximization algorithm to perform maximum a posteriori estimation to update the state estimate at each time step. Their modification makes it practical to perform the inference through calls to a physics simulation engine.