Papers › Learning from Spatio-temporal Correlation for Semi-Supervised LiDAR Semantic Segmentation

Learning from Spatio-temporal Correlation for Semi-Supervised LiDAR Semantic Segmentation

9 Oct 2024arXiv:2410.06893archive 2025-07-28

Seungho Lee, Hwijeong Lee, Hyunjung Shim

We address the challenges of the semi-supervised LiDAR segmentation (SSLS) problem, particularly in low-budget scenarios. The two main issues in low-budget SSLS are the poor-quality pseudo-labels for unlabeled data, and the performance drops due to the significant imbalance between ground-truth and pseudo-labels. This imbalance leads to a vicious training cycle. To overcome these challenges, we leverage the spatio-temporal prior by recognizing the substantial overlap between temporally adjacent LiDAR scans. We propose a proximity-based label estimation, which generates highly accurate pseudo-labels for unlabeled data by utilizing semantic consistency with adjacent labeled data. Additionally, we enhance this method by progressively expanding the pseudo-labels from the nearest unlabeled scans, which helps significantly reduce errors linked to dynamic classes. Additionally, we employ a dual-branch structure to mitigate performance degradation caused by data imbalance. Experimental results demonstrate remarkable performance in low-budget settings (i.e., <= 5%) and meaningful improvements in normal budget settings (i.e., 5 - 50%). Finally, our method has achieved new state-of-the-art results on SemanticKITTI and nuScenes in semi-supervised LiDAR segmentation. With only 5% labeled data, it offers competitive results against fully-supervised counterparts. Moreover, it surpasses the performance of the previous state-of-the-art at 100% labeled data (75.2%) using only 20% of labeled data (76.0%) on nuScenes. The code is available on https://github.com/halbielee/PLE.

PaperPDFCode

Code

halbielee/ple officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

LIDAR Semantic SegmentationSemantic SegmentationSemi-Supervised Semantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (Voxel) mIoU (0.5% Labels) 52.2 #1 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (Voxel) mIoU (1% Labels) 61.1 #1 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (Voxel) mIoU (10% Labels) 63.1 #1 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (Voxel) mIoU (2% Labels) 62.9 #1 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (Voxel) mIoU (20% Labels) 64.1 #1 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (Voxel) mIoU (5% Labels) 62.8 #1 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (Voxel) mIoU (50% Labels) 64.3 #1 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI LaserMix (Voxel) mIoU (0.5% Labels) 47.3 #2 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI LaserMix (Voxel) mIoU (2% Labels) 59.2 #2 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI LaserMix (Voxel) mIoU (5% Labels) 61.7 #2 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (CENet, Range view) mIoU (0.5% Labels) 46.2 #3 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (CENet, Range view) mIoU (1% Labels) 51.5 #3 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (CENet, Range view) mIoU (2% Labels) 54.3 #3 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation SemanticKITTI PLE (CENet, Range view) mIoU (5% Labels) 58.1 #3 of 12 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes PLE (Voxel) mIoU (0.5% Labels) 58 #1 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes PLE (Voxel) mIoU (1% Labels) 62.9 #1 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes PLE (Voxel) mIoU (10% Labels) 74.3 #1 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes PLE (Voxel) mIoU (2% Labels) 67.2 #1 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes PLE (Voxel) mIoU (20% Labels) 76 #1 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes PLE (Voxel) mIoU (5% Labels) 72.8 #1 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes PLE (Voxel) mIoU (50% Labels) 76.1 #1 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes LaserMix (Voxel) mIoU (0.5% Labels) 51.4 #2 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes LaserMix (Voxel) mIoU (2% Labels) 63.9 #2 of 11 Archive leaderboard report
Semi-Supervised Semantic Segmentation nuScenes LaserMix (Voxel) mIoU (5% Labels) 69.7 #2 of 11 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections