Collaborative Learning for Semi-Supervised LiDAR Semantic Segmentation

TL;DR

The paper introduces CoLLiS, a collaborative learning framework for semi-supervised LiDAR segmentation that mitigates confirmation bias by training multiple representations simultaneously and adaptively distilling knowledge.

Contribution

CoLLiS is a novel framework that trains multiple LiDAR representations collaboratively in a single step, reducing bias and improving semi-supervised segmentation performance.

Findings

CoLLiS outperforms state-of-the-art methods on three datasets.

Significant gains in low-label regimes.

Effective mitigation of confirmation bias during training.

Abstract

Annotating large-scale LiDAR point clouds for 3D semantic segmentation is costly and time-consuming, which motivates the use of semi-supervised learning (SemiSL). Standard LiDAR SemiSL methods typically adopt a two-step training paradigm, where pseudo-labels are separately generated from a single distillation source, either from the same or another LiDAR representation. Such supervision relies on a unique source of pseudo-labels, which can reinforce confirmation bias and propagate errors during training, ultimately limiting performance. To address this challenge, we introduce CoLLiS, a novel framework that leverages Collaborative Learning for LiDAR Semi-supervised segmentation. Unlike prior paradigms with decoupled pseudo-labeling and training phases, CoLLiS trains multiple representations collaboratively in a single step by treating them as coequal students. Each student is adaptively distilled from multiple representations, while inter-student disparities are monitored online to resolve contradictory supervision and effectively mitigate confirmation bias. Extensive experiments on three datasets demonstrate that CoLLiS consistently outperforms state-of-the-art LiDAR SemiSL methods, with particularly strong gains in low-label regimes.