Research map / Detection and segmentation / CNN features
DAMap: Distance-aware MapNet for High Quality HD Map Construction
Jinpeng Dong, Chen Li, Yutong Lin, Jingwen Fu, Sanping Zhou, Nanning Zheng
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 DAMap (DAFL+HLS+TMDA) | MapTRv2 · nuScenes val easy (MapTRv2, R50, 24 ep) | mAP: 60.4 -> 62.8 | Improved |
| add DAMap | MapQR · nuScenes val easy (MapTRv2, R50, 24 ep) | mAP: 66.4 -> 68.8 | Improved |
| add DAMap | Mask2Map · nuScenes val easy (MapTRv2, R50, 24 ep) | mAP: 71.6 -> 72.6 | Improved |
| add DAMap with SwinB backbone | MapQR · nuScenes val easy (SwinB) | mAP: 70.0 -> 71.9 | Improved |
| add DAMap on Argoverse2 (24 epochs) | MapTRv2 · Argoverse2 val easy | mAP: 68.5 -> 69.4 | Improved |
| add DAMap with centerline category | MapTRv2 · nuScenes val (4 classes incl. centerline) | mAP: 53.2 -> 58.3 | Improved |
| add Distance-aware Focal Loss (DAFL) (ablation) | MapTRv2 · nuScenes val easy | mAP: 60.4 -> 60.9 | Improved |
| add Hybrid Loss Scheme on top of DAFL (ablation) | MapTRv2+DAFL · nuScenes val easy | mAP: 60.9 -> 61.6 | Improved |
| add Task Modulated Deformable Attention (TMDA) (ablation) | MapTRv2 · nuScenes val easy | mAP: 60.4 -> 61.6 | Improved |
| add TMDA and HLS to DAFL (ablation) | MapTRv2+DAFL · nuScenes val easy | mAP: 61.6 -> 62.8 | Improved |
| use separate task-specific offsets and weights (TMDA design 1) instead of TMDA (ablation) | MapTRv2 · nuScenes val easy | mAP: 60.4 -> 60.7 | Improved |
| use TMDA with task-agnostic offsets and task-specific attention weights vs separate cross-attention (ablation) | MapTRv2 · nuScenes val easy | mAP: 60.7 -> 61.6 | 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.
- Mask2Map — Mask2Map: Vectorized HD Map Construction Using Bird's Eye View Segmentation Masks (ECCV 2024) · 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
- MGMap: Mask-Guided Learning for Online Vectorized HD Map Construction (CVPR 2024 · similarity 0.88)
- SDTagNet: Leveraging Text-Annotated Navigation Maps for Online HD Map Construction (NeurIPS 2025 · similarity 0.87)
- MapExpert: Online HD Map Construction with Simple and Efficient Sparse Map Element Expert (AAAI 2025 · similarity 0.87)
- Unveiling the Hidden: Online Vectorized HD Map Construction with Clip-Level Token Interaction and Propagation (NeurIPS 2024 · similarity 0.87)
- Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction (CVPR 2025 · similarity 0.87)
- HIMap: HybrId Representation Learning for End-to-end Vectorized HD Map Construction (CVPR 2024 · 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.
