3-D PersonVLAD: Learning Deep Global Representations for Video-Based Person Reidentification | IEEE Journals & Magazine | IEEE Xplore