3D Object Detection with a Deformable 3D Cuboid Model

This paper addresses the problem of category-level 3D object detection. Given a monocular image, our aim is to localize the objects in 3D by enclosing them with tight oriented 3D bounding boxes. We propose a novel approach that extends the deformable part-based model [1] to reason in 3D. Our model represents an object class as a deformable 3D cuboid composed of faces and parts, which are both allowed to deform with respect to their anchors on the 3D box. We model the appearance of each face in fronto-parallel coordinates, thus effectively factoring out the appearance variation induced by viewpoint. We train the cuboid model jointly and discriminatively. In inference we slide and rotate the box in 3D to score the object hypotheses. We evaluate our approach in indoor and outdoor scenarios, and show that our approach outperforms the state-of-the-art in both 2D [1] and 3D object detection [4].