Mining discriminative 3D Poselet for cross-view action recognition | IEEE Conference Publication | IEEE Xplore