RePL: Pseudo-label Refinement for Semi-supervised LiDAR Semantic Segmentation

Donghyeon Kwon; Taegyu Park; Suha Kwak

cs.CVApr 8, 2026

RePL: Pseudo-label Refinement for Semi-supervised LiDAR Semantic Segmentation

Authors: Donghyeon Kwon, Taegyu Park, Suha Kwak

Organizations: GENGENAI, Republic of Korea · Dept. of CSE, POSTECH, Republic of Korea · Graduate School of AI, POSTECH, Republic of Korea

Abstract

Semi-supervised learning for LiDAR semantic segmentation often suffers from error propagation and confirmation bias caused by noisy pseudo-labels. To tackle this chronic issue, we introduce RePL, a novel framework that enhances pseudo-label quality by identifying and correcting potential errors in pseudo-labels through masked reconstruction, along with a dedicated training strategy. We also provide a theoretical analysis demonstrating the condition under which the pseudo-label refinement is beneficial, and empirically confirm that the condition is mild and clearly met by RePL. Extensive evaluations on the nuScenes-lidarseg and SemanticKITTI datasets show that RePL improves pseudo-label quality substantially, and in consequence, achieves the state of the art in semi-supervised LiDAR semantic segmentation.

Figures & tables

Figure 1: Overview of RePL . The teacher generates predictions for unlabeled LiDAR scenes, which are used as pseudo-labels for the student, and is updated via exponential moving average (EMA) of the student. The pseudo-label refiner detects erroneous pseudo-labels by confidence-based agreement between the teacher and student, and then corrects them through masked reconstruction with learnable tokens. The final refined pseudo-labels combine reliable teacher predictions with reconstructed ones, yielding improved supervision for semi-supervised learning of the student.

Panel of Figure 2 — Figure 2: Visualization of the improvement condition from Eq. 11 on the SemanticKITTI dataset given $\pi$ . The two values of $\pi$ were derived during the actual experiments on the dataset: 1% labeled data ( $\pi=0.917$ ) and 50% labeled data ( $\pi=0.983$ ). Green areas indicate combinations of $q$ and $r$ where the refinement yields a net benefit ( $\zeta>0$ ), while red areas show such combinations leading to detriment ( $\zeta\leq 0$ ). The results suggest that the refinement remains beneficial across a broad range of $q$ and $r$ , and that RePL clearly helps improve the quality of pseudo-labels.

Method	Backbone	nuScenes-lidarseg [ 6 ]					SemanticKITTI [ 1 ]
Method		1%	10%	20%	50%	Avg.	1%	10%	20%	50%	Avg.
Sup-only	Cylinder3D	50.9	65.9	66.6	71.2	63.7	45.4	56.1	57.8	58.7	54.5
Seal [ 23 ]	MinkUNet ${}^{\text{*}}$	45.8	63.0	-	-	-	46.6	-	-	-	-
SuperFlow [ 40 ]	MinkUNet ${}^{\text{*}}$	48.1	64.5	-	-	-	48.4	-	-	-	-
SLidR [ 32 ]	Cylinder3D ${}^{\text{*}}$	39.0	58.8	-	-	-	44.6	-	-	-	-
Lim3D [ 20 ]	LiM3D	-	-	-	-	-	58.4	59.5	63.1	63.6	61.2

Table 1: Comparison of different semi-supervised learning methods on nuScenes-lidarseg and SemanticKITTI while varying the ratio of labeled data. The best results in each column are shown in bold, and the second best are underlined . Backbones marked with an asterisk (*) indicate that additional representation learning or knowledge distillation from external sources has been applied.

Table 4

Table 5

Panel of Figure 3 — Figure 3: Qualitative results of refined pseudo-labels and their initial predictions on the unlabeled set of nuScenes-lidarseg, at the end of training. Correct and incorrect predictions are shown in green and red, respectively. The model was trained with a 1% label ratio.

Table 7

Panel of Figure 4 — Figure 4: Failure cases on the unlabeled set of nuScenes-lidarseg at the end of training with a 1% label ratio. Correct and incorrect predictions are shown in green and red, respectively.

Figure 5: Pseudo-label quality improvement by the refiner during training on nuScenes-lidarseg.

Figure 6: Sensitivity analysis of a hyper-parameter

\kappa

Class	%	$\Delta$ IoU	Class	%	$\Delta$ IoU	Class	%	$\Delta$ IoU	Class	%	$\Delta$ IoU
barrier	1.10	+3.45	bus	0.55	+1.04	const. veh.	0.18	+2.54	motorcycle	0.05	+2.55
bicycle	0.02	+2.68	car	4.42	+1.60	pedestrian	0.27	+6.47	traffic cone	0.09	+7.86
trailer	0.58	+0.60	truck	1.83	+1.02	driv. surf.	37.6	$-$ 1.47	flat other	1.02	+2.60
sidewalk	8.31	$-$ 2.40	terrain	8.34	+4.14	manmade	21.1	+4.41	vegetation	14.5	+5.35

Table 8: Per-class pseudo-label quality improvement on the unlabeled set of nuScenes-lidarseg with 1% labeled ratio. The column marked as % indicates class frequency.

Semantic Segmentation

Semantic segmentation assigns a class label to every image pixel, enabling detailed scene understanding across domains and sensors.

See all 414 papers

Collaborative Learning for Semi-Supervised LiDAR Semantic Segmentation
May 16, 2026Bin Yang, Alexandru Paul Condurache3D Semantic Segmentation Semi-Supervised Learning
Bosch Research, Robert Bosch GmbH, Stuttgart, Germany · Institute for Neuro- and Bioinformatics, University of Lübeck, Lübeck, Germany
Exploring Easy Boosts for Lidar Semantic Scene Completion
Jun 2, 2026Tetiana Martyniuk, Jonathan Seele, Alexandre Boulch +33D Semantic Segmentation Pseudo-Labeling
Inria, France · ETH Zurich, Switzerland · valeo.ai, France +1
Unsupervised Point Cloud Registration via Training-Time Semantic Guidance
Sep 14, 2026Kezheng Xiong, Shiyun Xu, Sheng Ao +3Point Cloud Registration
Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, Xiamen University, China · Fujian Key Laboratory of Urban Intelligent Sensing and Computing, Xiamen University, China

May 16, 2026J/K move · Enter open · S save

Collaborative Learning for Semi-Supervised LiDAR Semantic Segmentation

Bin Yang, Alexandru Paul Condurache

Bosch Research, Robert Bosch GmbH, Stuttgart, Germany · Institute for Neuro- and Bioinformatics, University of Lübeck, Lübeck, Germany

3D Semantic Segmentation Semi-Supervised Learning Knowledge Distillation

Open paper AskPDF arXiv

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.

Abstract

Semantic Segmentation

Explore similar work

Collaborative Learning for Semi-Supervised LiDAR Semantic Segmentation

Exploring Easy Boosts for Lidar Semantic Scene Completion

Unsupervised Point Cloud Registration via Training-Time Semantic Guidance