Research map / Detection and segmentation / CNN features
Benchmarking the Robustness of LiDAR Semantic Segmentation Models
Xu Yan, Chaoda Zheng, Ying Wen Xue, Zhen Li, Shuguang Robert Cui, Dengxin Dai
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 |
|---|---|---|---|
| use larger range image (2048x64 instead of 512x64) (ablation) | CENet · SemanticKITTI-C | mR: 70.4 -> 76.6 | Improved |
| use larger range image (2048x64 instead of 512x64) (ablation) | CENet · SemanticKITTI-C local distortion | mIoU: 50.5 -> 60.4 | Improved |
| use smaller range image (512x64 instead of 2048x64) for low-beam LiDAR (ablation) | CENet · SemanticKITTI-C 32-beam | R: 87.5 -> 90.0 | Improved |
| add attentive pooling to RandLA-Net (ablation) | RandLA-Net · SemanticKITTI-C | mR: 76.8 -> 80.4 | Improved |
| add attention pooling and relative-position MLP transformation to RandLA-Net (ablation) | RandLA-Net · SemanticKITTI-C | mR: 66.7 -> 80.4 | Improved |
| use cylindrical instead of grid voxel partition (ablation) | MinkowskiNet · SemanticKITTI-C fog | R: 84.9 -> 94.0 | Improved |
| use cylindrical instead of grid voxel partition (ablation) | MinkowskiNet · SemanticKITTI-C snowfall | R: 76.1 -> 93.0 | Improved |
| use cylindrical instead of grid voxel partition (ablation) | MinkowskiNet · SemanticKITTI-C 16-beam | R: 76.0 -> 50.3 | Hurt |
| use cylindrical instead of grid voxel partition (ablation) | MinkowskiNet · SemanticKITTI-C clean | mIoU: 66.3 -> 63.3 | Hurt |
| increase voxel size from 5cm to 20cm (ablation) | MinkowskiNet · SemanticKITTI-C global outliers | R: 98.5 -> 77.7 | Hurt |
| add PointHead to 2DPASS (ablation) | 2DPASS · SemanticKITTI-C | clean mIoU: 65.4 -> 70.1 | Improved |
| add PointHead and point branch (hybrid) to 2DPASS (ablation) | 2DPASS · SemanticKITTI-C | mR: 74.2 -> 72.9 | Hurt |
Closest papers by meaning
- Test-Time Training for LiDAR Semantic Segmentation under Corruption via Geometric Inlier Discrimination (CVPR 2026 · similarity 0.92)
- No Thing, Nothing: Highlighting Safety-Critical Classes for Robust LiDAR Semantic Segmentation in Adverse Weather (CVPR 2025 · similarity 0.91)
- Rethinking Data Augmentation for Robust LiDAR Semantic Segmentation in Adverse Weather (ECCV 2024 · similarity 0.90)
- Benchmarking the Robustness of Semantic Segmentation Models with Respect to Common Corruptions (IJCV 2020 · similarity 0.90)
- UniMix: Towards Domain Adaptive and Generalizable LiDAR Semantic Segmentation in Adverse Weather (CVPR 2024 · similarity 0.89)
- Multi-Space Alignments Towards Universal LiDAR Segmentation (CVPR 2024 · 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.
