Learning Inter-Superpoint Affinity for Weakly Supervised 3D Instance Segmentation

Linghua Tang, Le Hui, Jin Xie; Proceedings of the Asian Conference on Computer Vision (ACCV), 2022, pp. 1282-1297


Weakly supervised 3D instance segmentation on point clouds has been rarely studied in recent years. Due to the few annotated labels of 3D point clouds, how to learn discriminative features of point clouds to segment object instances is a challenging problem. In this paper, we propose a simple but effective instance segmentation framework that can achieve striking performance by annotating only one point for each instance. Specifically, to tackle extremely few labels, we first oversegment the point cloud into superpoints in an unsupervised manner and extend the point-level annotations to the superpoint level. Then, based on the superpoint graph, we propose an inter-superpoint affinity mining module that considers the semantic and spatial relations to adaptively learns inter-superpoint affinity to generate high-quality pseudo labels via random walk. Finally, we propose a volume-aware instance refinement module to segment high-quality instances by applying volume constraints of objects in clustering on the superpoint graph. Extensive experiments on the ScanNet-v2 and S3DIS datasets demonstrate that our method achieves state-of-the-art performance in the weakly supervised point cloud instance segmentation task, and even outperforms some fully supervised methods.

Related Material

[pdf] [supp] [arXiv] [code]
@InProceedings{Tang_2022_ACCV, author = {Tang, Linghua and Hui, Le and Xie, Jin}, title = {Learning Inter-Superpoint Affinity for Weakly Supervised 3D Instance Segmentation}, booktitle = {Proceedings of the Asian Conference on Computer Vision (ACCV)}, month = {December}, year = {2022}, pages = {1282-1297} }