Learning Compact Hash Codes for Multimodal Representations Using Orthogonal Deep Structure | IEEE Journals & Magazine | IEEE Xplore