Collaborative Learning for Semi-Supervised LiDAR Semantic Segmentation

2026-05-16

AI Summary

This work addresses the high annotation cost in LiDAR point cloud 3D semantic segmentation and the confirmation bias and error propagation inherent in existing semi-supervised methods that rely on a single pseudo-label source. To overcome these limitations, we propose CoLLiS, a novel framework that departs from the conventional two-stage pseudo-labeling paradigm by introducing a single-stage, multi-representation collaborative learning mechanism. In CoLLiS, multiple LiDAR representations act as peer students that mutually learn from one another through adaptive knowledge distillation, while an online discrepancy monitoring module resolves conflicting supervisory signals. Extensive experiments demonstrate that CoLLiS significantly outperforms current semi-supervised approaches across three standard benchmarks, achieving particularly remarkable gains under extremely low labeling rates.

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.

Problem

• LiDAR semantic segmentation
• semi-supervised learning
• pseudo-labeling
• confirmation bias
• 3D point clouds

Innovation

• Collaborative Learning
• Semi-supervised Learning
• LiDAR Semantic Segmentation
• Pseudo-labeling
• Confirmation Bias Mitigation