View-invariant representation and learning of human action | IEEE Conference Publication | IEEE Xplore