Three-dimensional objects are commonly represented as 3D boxes in a point-cloud. This representation mimics the well-studied image-based 2D bounding-box detection but comes with additional challenges. Objects in a 3D world do not follow any particular orientation, and box-based detectors have difficulties enumerating all orientations or fitting an axis-aligned bounding box to rotated objects. In this paper, we instead propose to represent, detect, and track 3D objects as points. We use a keypoint detector to find centers of objects and simply regress to other attributes, including 3D size, 3D orientation, and velocity. In our center-based framework, 3D object tracking simplifies to greedy closest-point matching. The resulting detection and tracking algorithm is simple, efficient, and effective. On the nuScenes dataset, our point-based representations perform $3$-$4$ mAP higher than the box-based counterparts for 3D detection, and 6 AMOTA higher for 3D tracking. Our real-time model runs end-to-end 3D detection and tracking at $30$ FPS with $54.2$ AMOTA and $48.3$ mAP while the best single model achieves $60.3$ mAP for 3D detection and $63.8$ AMOTA for 3D tracking. The code and pretrained models are available at https://github.com/tianweiy/CenterPoint.
URL: http://arxiv.org/abs/2006.11275v1