View-Invariant Human Action Recognition Via View Transformation Network (VTN) | IEEE Journals & Magazine | IEEE Xplore