End-to-end 3D point cloud instance segmentation without detection

3D instance segmentation plays a predominant role in environment perception of robotics and augmented reality. Many deep learning based methods have been presented recently for this task. These methods rely on either a detection branch to propose objects or a grouping step to assemble same-instance points. However, detection based methods do not ensure a consistent instance label for each point, while the grouping step requires parameter-tuning and is computationally expensive. In this paper, we introduce an assign-and-suppress network, dubbed as AS-Net, to enable end-to-end instance segmentation without detection and a separate step of grouping. The core idea is to frame instance segmentation as a candidate assignment problem. At first, a set of instance candidates are sampled. Then we propose an assignment module for candidate assignment and a suppression module to eliminate redundant candidates. A mapping between instance labels and instance candidates is further sought to construct an instance grouping loss for the network training. Experimental results demonstrate that our method is more effective and efficient than previous detection-free approaches.