Research map / Point clouds / Segmentation and detection
PolarMix: A General Data Augmentation Technique for LiDAR Point Clouds
Aoran Xiao, Jiaxing Huang, Dayan Guan, Kaiwen Cui, Shijian Lu, Ling Shao
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 PolarMix to MinkNet training | MinkNet · SemanticKITTI val | mIoU: 55.9 -> 65.0 | Improved |
| Add PolarMix to SPVCNN training | SPVCNN · SemanticKITTI val | mIoU: 58.0 -> 66.2 | Improved |
| Add PolarMix to MinkNet training | MinkNet · nuScenes-lidarseg val | mIoU: 67.1 -> 72.0 | Improved |
| Add PolarMix to MinkNet training | MinkNet · SemanticPOSS val | mIoU: 52.1 -> 57.4 | Improved |
| Add PolarMix to SPVCNN training | SPVCNN · nuScenes-lidarseg val | mIoU: 68.4 -> 72.1 | Improved |
| Add PolarMix to SPVCNN training | SPVCNN · SemanticPOSS val | mIoU: 50.7 -> 58.6 | Improved |
| Add PolarMix | RandLA-Net · SemanticKITTI val (10% training data) | mIoU: 45.4 -> 50.5 | Improved |
| Add PolarMix | Cylinder3D · SemanticKITTI val (10% training data) | mIoU: 60.6 -> 62.5 | Improved |
| Add PolarMix augmentation | PointPillar, Second, and CenterNet · nuScenes val | mAP: 41.8 -> 43.7; 48.8 -> 50.3; 52.8 -> 55.4 | Improved |
| Mix labeled source scans with pseudo-labeled target scans using PolarMix | MinkNet self-training · SynLiDAR → SemanticKITTI | mIoU: 28.9 -> 31.0 | Improved |
| Mix labeled source scans with pseudo-labeled target scans using PolarMix | MinkNet self-training · SynLiDAR → SemanticPOSS | mIoU: 29.6 -> 30.4 | Improved |
| Add scene-level swapping (ablation) | SPVCNN · SemanticKITTI val (sequence 00 training) | mIoU: 48.9 -> 50.8 | Improved |
Later papers that compared with this method
1 later accepted paper reports PolarMix as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- 3D Semantic Segmentation in the Wild: Learning Generalized Models for Adverse-Condition Point Clouds (CVPR 2023 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- RandLA-Net — RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds (CVPR 2020) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- BeyondMix: Leveraging Structural Priors and Long-Range Dependencies for Domain-Invariant LiDAR Segmentation (NeurIPS 2025 · similarity 0.88)
- RPVNet: A Deep and Efficient Range-Point-Voxel Fusion Network for LiDAR Point Cloud Segmentation (ICCV 2021 · similarity 0.88)
- Not All Points Are Equal: Learning Highly Efficient Point-Based Detectors for 3D LiDAR Point Clouds (CVPR 2022 · similarity 0.88)
- LaserMix for Semi-Supervised LiDAR Semantic Segmentation (CVPR 2023 · similarity 0.87)
- LiDAR-Net: A Real-scanned 3D Point Cloud Dataset for Indoor Scenes (CVPR 2024 · similarity 0.87)
- Unlocking Generalization Power in LiDAR Point Cloud Registration (CVPR 2025 · similarity 0.86)
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.
