Research map / Detection and segmentation / 3D detection and driving
VectorMapNet: End-to-end Vectorized HD Map Learning
Yicheng Liu, Tianyuan Yuan, Yue Wang, Yilun Wang, Hang Zhao
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 VectorMapNet instead of HDMapNet in fusion setting | nuScenes map learning · nuScenes | mAP: 31.0 -> 45.2 | Improved |
| Fine-tune using predicted keypoints after teacher forcing (ablation) | VectorMapNet (Camera) · nuScenes | mAP: +6.9 mAP | Improved |
| Fine-tune using predicted keypoints after teacher forcing (ablation) | VectorMapNet (Fusion) · nuScenes | mAP: +8.5 mAP | Improved |
| Fuse camera and LiDAR inputs instead of using camera alone | VectorMapNet · nuScenes | mAP: +4.3 mAP | Improved |
| Fuse camera and LiDAR inputs instead of using LiDAR alone | VectorMapNet · nuScenes | mAP: +11.2 mAP | Improved |
| Use Bbox keypoints instead of SME keypoints (ablation) | VectorMapNet · Argoverse2 | Fréchet mAP: 50.4 -> 52.4 | Improved |
| Use Bbox keypoints instead of SME keypoints (ablation) | VectorMapNet · Argoverse2 | Chamfer mAP: 32.5 -> 40.9 | Improved |
| Replace curvature-based sampling with fixed-interval sampling (ablation) | VectorMapNet polyline generator · nuScenes | Fréchet mAP: 50.4 -> 28.9 | Hurt |
| Replace curvature-based sampling with fixed-interval sampling (ablation) | VectorMapNet polyline generator · nuScenes | Chamfer mAP: 32.5 -> 17.0 | Hurt |
| Replace discrete vertex modeling with continuous mixture-of-Gaussians modeling (ablation) | VectorMapNet polyline generator · nuScenes | Fréchet mAP: 41.9 -> 50.4 | Hurt |
| Replace discrete vertex modeling with continuous mixture-of-Gaussians modeling (ablation) | VectorMapNet polyline generator · nuScenes | Chamfer mAP: 26.5 -> 32.5 | Hurt |
| Add map predicted by VectorMapNet to trajectory inputs | mmTransformer · nuScenes motion forecasting | minADE: 0.909 -> 0.826 | Improved |
Closest papers by meaning
- InteractionMap: Improving Online Vectorized HDMap Construction with Interaction (CVPR 2025 · similarity 0.90)
- SDTagNet: Leveraging Text-Annotated Navigation Maps for Online HD Map Construction (NeurIPS 2025 · similarity 0.90)
- Unveiling the Hidden: Online Vectorized HD Map Construction with Clip-Level Token Interaction and Propagation (NeurIPS 2024 · similarity 0.89)
- MapExpert: Online HD Map Construction with Simple and Efficient Sparse Map Element Expert (AAAI 2025 · similarity 0.89)
- Driving by the Rules: A Benchmark for Integrating Traffic Sign Regulations into Vectorized HD Map (CVPR 2025 · similarity 0.88)
- Patched Line Segment Learning for Vector Road Mapping (AAAI 2024 · 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.
