Research map / Point clouds / Segmentation and detection
Rethinking Range View Representation for LiDAR Segmentation
Lingdong Kong, Youquan Liu, Runnan Chen, Yuexin Ma, Xinge Zhu, Yikang Li, Yuenan Hou, Yu Qiao, Ziwei Liu
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 |
|---|---|---|---|
| Cityscapes pre-training for RangeFormer (ablation) | RangeFormer · SemanticKITTI val (64x2048) | mIoU: 68.1 -> 69.6 | Improved |
| ImageNet pre-training for FIDNet (ablation) | FIDNet · SemanticKITTI val (64x2048) | mIoU: 60.4 -> 61.6 | Improved |
| ImageNet pre-training for CENet (ablation) | CENet · SemanticKITTI val (64x2048) | mIoU: 63.4 -> 64.1 | Improved |
| Cityscapes pre-training for RangeFormer (ablation) | RangeFormer · nuScenes val | mIoU: 77.1 -> 78.1 | Improved |
| train with STR multi-view partition (W_train=384) instead of W=2048 | CENet · SemanticKITTI test | mIoU: 64.7 -> 65.8 | Improved |
| train with STR instead of W=2048 | FIDNet · SemanticKITTI test | mIoU: 59.5 -> 60.1 | Improved |
| train RangeFormer with STR instead of W=2048 | RangeFormer · SemanticKITTI test | mIoU: 73.3 -> 72.2 | Mixed |
| split into more than 5 views in STR (ablation) | STR · SemanticKITTI val | mIoU | Hurt |
| apply RangeAug augmentation instead of common augmentation (ablation) | RangeFormer, CENet, FIDNet · SemanticKITTI val | mIoU | Improved |
| apply RangePost supervised sub-cloud post-processing instead of k-NN or NLA (ablation) | range view segmenters · SemanticKITTI val | mIoU | Improved |
| train RangeFormer without data augmentation (plain) (ablation) | RangeFormer · SemanticKITTI val | mIoU | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- 2DPASS — 2DPASS: 2D Priors Assisted Semantic Segmentation on LiDAR Point Clouds (ECCV 2022) · better on 1 of 1 shared result
- RPVNet — RPVNet: A Deep and Efficient Range-Point-Voxel Fusion Network for LiDAR Point Cloud Segmentation (ICCV 2021) · 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
- RangeViT: Towards Vision Transformers for 3D Semantic Segmentation in Autonomous Driving (CVPR 2023 · similarity 0.92)
- RAPiD-Seg: Range-Aware Pointwise Distance Distribution Networks for 3D LiDAR Segmentation (ECCV 2024 · similarity 0.90)
- Spiral: Semantic-Aware Progressive LiDAR Scene Generation and Understanding (NeurIPS 2025 · similarity 0.89)
- Multi-Space Alignments Towards Universal LiDAR Segmentation (CVPR 2024 · similarity 0.89)
- PC-BEV: An Efficient Polar-Cartesian BEV Fusion Framework for LiDAR Semantic Segmentation (AAAI 2025 · similarity 0.89)
- TREND: Unsupervised 3D Representation Learning via Temporal Forecasting for LiDAR Perception (NeurIPS 2025 · similarity 0.88)
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.
