We propose a novel, conceptually simple and general framework for instance segmentation on 3D point clouds. Our method, called 3D-BoNet, follows the simple design philosophy of per-point multilayer perceptrons (MLPs). The framework directly regresses 3D bounding boxes for all instances in a point cloud, while simultaneously predicting a point-level mask for each instance. It consists of a backbone network followed by two parallel network branches for 1) bounding box regression and 2) point mask prediction. 3D-BoNet is single-stage, anchor-free and end-to-end trainable. Moreover, it is remarkably computationally efficient as, unlike existing approaches, it does not require any post-processing steps such as non-maximum suppression, feature sampling, clustering or voting. Extensive experiments show that our approach surpasses existing work on both ScanNet and S3DIS datasets while being approximately 10× more computationally efficient. Comprehensive ablation studies demonstrate the effectiveness of our design.
|Title of host publication||Advances in Neural Information Processing Systems (NIPS 2019)|
|Publisher||Neural Information Processing Systems Foundation|
|Number of pages||11|
|Publication status||Accepted/In press - 4 Sep 2019|
|Event||2019 Conference on Neural Information Processing Systems - Vancouver, Canada|
Duration: 8 Dec 2019 → 14 Dec 2019
|Conference||2019 Conference on Neural Information Processing Systems|
|Abbreviated title||NeurIPS 2019|
|Period||8/12/19 → 14/12/19|
Yang, B., Wang, J., Clark, R., Hu, Q., Wang, S., Markham, A., & Trigoni, N. (Accepted/In press). Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds. In Advances in Neural Information Processing Systems (NIPS 2019) (pp. 1-10). Neural Information Processing Systems Foundation.