Research map / Label-efficient and robust learning / Semi-supervised learning
ItTakesTwo: Leveraging Peer Representations for Semi-supervised LiDAR Semantic Segmentation
Yuyuan Liu, Yuanhong Chen, Hu Wang, Vasileios Belagiannis, Ian Reid, Gustavo Carneiro
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add peer-representation consistency (IT2 architecture) in place of single-representation CPS (ablation) | IT2 · nuScenes 10% (range) | mIoU: 60.8 -> 67.3 | Improved |
| add peer-representation consistency in place of single-representation CPS (ablation) | IT2 · nuScenes 10% (voxel) | mIoU: 66.3 -> 70.4 | Improved |
| add GMM-based cross-distribution contrastive learning (ablation) | IT2 · nuScenes 10% (range) | mIoU: 67.3 -> 70.3 | Improved |
| add representation-specific augmentation (multi-box CutMix for range, single-inclination LaserMix for voxel) (ablation) | IT2 · ScribbleKITTI 10% (range) | mIoU: 54.9 -> 57.1 | Improved |
| add representation-specific augmentation (ablation) | IT2 · nuScenes 50% (voxel) | mIoU: 73.9 -> 74.1 | No clear effect |
| use GMM prototype sampling instead of label-guided pixel-wise contrastive learning | IT2 · nuScenes | mIoU | Improved |
| use IT2 versus LaserMix at 50% labels on ScribbleKITTI range | IT2 · ScribbleKITTI 50% (range) | mIoU: 58.7 -> 58.6 | No clear effect |
| use test-time augmentation with IT2 | IT2 · SemanticKITTI partial 20% | mIoU: 56.5 -> 58.8 | Improved |
| add a BEV representation as peer to range or voxel (ablation) | IT2 · nuScenes 10% | mIoU: +2.4 / +3.9 mIoU | Improved |
Closest papers by meaning
- Collaborative Learning for Semi-Supervised LiDAR Semantic Segmentation (ICML 2026 · similarity 0.92)
- LaserMix for Semi-Supervised LiDAR Semantic Segmentation (CVPR 2023 · similarity 0.90)
- RePL: Pseudo-label Refinement for Semi-supervised LiDAR Semantic Segmentation (ECCV 2026 · similarity 0.89)
- On-the-fly Category Discovery for LiDAR Semantic Segmentation (ECCV 2024 · similarity 0.88)
- Exploring Scene Affinity for Semi-Supervised LiDAR Semantic Segmentation (CVPR 2025 · similarity 0.88)
- PC-BEV: An Efficient Polar-Cartesian BEV Fusion Framework for LiDAR Semantic Segmentation (AAAI 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
