Class agnostic and specific consistency learning for weakly-supervised point cloud semantic segmentation

Junwei Wu, Mingjie Sun, Haotian Xu, Chenru Jiang, Wuwei Ma, Quan Zhang

Abstract: Highlights•Propose a weakly-supervised point cloud segmentation under 0.1% and 0.01% annotation.•Pull group-to-group point features closer using nearby perturbed point features.•steer unlabeled points towards corresponding class-specific memory bank features.