Shape and Symmetry Induction for 3D Objects
This work addresses the challenge of 3D shape understanding for robotics and vision tasks, such as grasping and navigation, but is incremental as it repurposes existing classification machinery for a new application.
The paper tackles the problem of recovering 3D shape from a single viewpoint for potentially unfamiliar objects by predicting surface normals and detecting symmetry planes, enabling extrapolation of occluded surfaces. It demonstrates accurate 3D shape recovery on object classes not seen during training, with results shown on both synthetic and real images.
Actions as simple as grasping an object or navigating around it require a rich understanding of that object's 3D shape from a given viewpoint. In this paper we repurpose powerful learning machinery, originally developed for object classification, to discover image cues relevant for recovering the 3D shape of potentially unfamiliar objects. We cast the problem as one of local prediction of surface normals and global detection of 3D reflection symmetry planes, which open the door for extrapolating occluded surfaces from visible ones. We demonstrate that our method is able to recover accurate 3D shape information for classes of objects it was not trained on, in both synthetic and real images.