Text this: Volumetric spatial feature representation for view-invariant human action recognition using a depth camera.