Multimodal Deep Autoencoder for Human Pose Recovery | IEEE Journals & Magazine | IEEE Xplore