Learning Viewpoint-Agnostic Visual Representations by Recovering Tokens in 3D Space

Open in new window