Deeply Learned View-Invariant Features for Cross-View Action Recognition | IEEE Journals & Magazine | IEEE Xplore