Research map / Point clouds / Segmentation and detection
LiDAR2Map: In Defense of LiDAR-Based Semantic Map Construction Using Online Camera Distillation
Song Wang, Wentong Li, Wenyu Liu, Xiaolu Liu, Jianke Zhu
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 BEV-FPD to the LiDAR map segmentation baseline | LiDAR2Map with PointPillars · nuScenes val, map segmentation | mIoU: 29.5% mIoU to 57.4% | Improved |
| Increase BEV-FPD from 2 to 4 layers (ablation) | LiDAR2Map with PointPillars · nuScenes val, map segmentation | mIoU: 43.8% mIoU; +10.5% mIoU | Improved |
| Increase BEV-FPD from 4 to 6 layers (ablation) | LiDAR2Map with PointPillars · nuScenes val, map segmentation | mIoU: 54.3; 57.4 | Improved |
| Add PGF²M (ablation) | LiDAR2Map · nuScenes val, map segmentation | mIoU: 52.2; 52.6 | Improved |
| Add PGF²M (ablation) | LiDAR2Map · nuScenes val, vehicle segmentation, Setting 2 | mIoU: 49.1; 50.6 | Improved |
| Add feature-level distillation to PGF²M (ablation) | LiDAR2Map · nuScenes val, map segmentation | mIoU: 52.6; 53.5 | Improved |
| Add feature-level distillation to PGF²M (ablation) | LiDAR2Map · nuScenes val, vehicle segmentation, Setting 2 | mIoU: 50.6; 51.8 | Improved |
| Add logit-level distillation to PGF²M (ablation) | LiDAR2Map · nuScenes val, map segmentation | mIoU: 52.6; 53.7 | Improved |
| Add logit-level distillation to PGF²M (ablation) | LiDAR2Map · nuScenes val, vehicle segmentation, Setting 2 | mIoU: 50.6; 51.3 | Improved |
| Use full online camera-to-LiDAR distillation (ablation) | LiDAR2Map · nuScenes val, map segmentation | mIoU: 52.2; 54.3 | Improved |
| Replace the proposed distillation scheme with MonoDistill or MGD (ablation) | LiDAR2Map · nuScenes val, map segmentation | mIoU: 52.2; 44.6; 50.1 | Hurt |
| Replace the proposed distillation scheme with xMUDA or 2DPASS (ablation) | LiDAR2Map · nuScenes val, map segmentation | mIoU: 52.2; 53.3; 53.6 | Improved |
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.
- 2DPASS — 2DPASS: 2D Priors Assisted Semantic Segmentation on LiDAR Point Clouds (ECCV 2022) · 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
- MapDistill: Boosting Efficient Camera-based HD Map Construction via Camera-LiDAR Fusion Model Distillation (ECCV 2024 · similarity 0.90)
- Map It Anywhere: Empowering BEV Map Prediction using Large-scale Public Datasets (NeurIPS 2024 · similarity 0.88)
- PC-BEV: An Efficient Polar-Cartesian BEV Fusion Framework for LiDAR Semantic Segmentation (AAAI 2025 · similarity 0.88)
- Spiral: Semantic-Aware Progressive LiDAR Scene Generation and Understanding (NeurIPS 2025 · similarity 0.88)
- VLM2Scene: Self-Supervised Image-Text-LiDAR Learning with Foundation Models for Autonomous Driving Scene Understanding (AAAI 2024 · similarity 0.88)
- SDGOCC: Semantic and Depth-Guided Bird's-Eye View Transformation for 3D Multimodal Occupancy Prediction (CVPR 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.
