3D Hierarchical Refinement and Augmentation for Unsupervised Learning of Depth and Pose from Monocular Video - 42Papers