View-Invariant 3D Action Recognition Using Spatiotemporal Self-Similarities from Depth Camera | IEEE Conference Publication | IEEE Xplore