Weakly Supervised Point Cloud Segmentation via Conservative Propagation of Scene-level Labels

TL;DR

This paper introduces a weakly supervised point cloud segmentation method that conservatively propagates scene-level labels to relevant points using primitive clustering and bipartite matching, significantly improving accuracy.

Contribution

The novel primitive-based label propagation approach reduces negative label influence and enhances segmentation performance under weak supervision.

Findings

Outperforms state-of-the-art on ScanNet and S3DIS datasets.

Effective label propagation reduces mislabeling in weak supervision.

Significant accuracy gains demonstrate the method's robustness.

Abstract

We propose a weakly supervised semantic segmentation method for point clouds that predicts "per-point" labels from just "whole-scene" annotations. The key challenge here is the discrepancy between the target of dense per-point semantic prediction and training losses derived from only scene-level labels. To address this, in addition to the typical weakly-supervised setup that supervises all points with the scene label, we propose to conservatively propagate the scene-level labels to points selectively. Specifically, we over-segment point cloud features via unsupervised clustering in the entire dataset and form primitives. We then associate scene-level labels with primitives through bipartite matching. Then, we allow labels to pass through this primitive-label relationship, while further encouraging features to form narrow clusters around the primitives. Importantly, through bipartite matching, this additional pathway through which labels flow, only propagates scene labels to the most relevant points, reducing the potential negative impact caused by the global approach that existing methods take. We evaluate our method on ScanNet and S3DIS datasets, outperforming the state of the art by a large margin.