Spatio-Temporal Video Analysis and the 3D Shearlet Transform