Research map / Video / Video segmentation and tracking
Stream Query Denoising for Vectorized HD-Map Construction
Shuo Wang, Fan Jia, Weixin Mao, Yingfei Liu, Yucheng Zhao, Zehui Chen, Tiancai Wang, Chi Zhang, Xiangyu Zhang, Feng 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 |
|---|---|---|---|
| add stream query denoising (SQD) training strategy (ablation) | StreamMapNet · nuScenes 30m | mAP: 59.2 -> 63.9 | Improved |
| add SQD to temporal StreamMapNet | StreamMapNet · nuScenes 30m | mAP: 60.4 -> 63.9 | Improved |
| add SQD to static MapTR | MapTR · nuScenes 30m | mAP: 50.3 -> 52.2 | Improved |
| add SQD to static BeMapNet | BeMapNet · nuScenes 30m | mAP: 59.8 -> 61.9 | Improved |
| use adaptive temporal matching instead of fixed threshold (ablation) | SQD-MapNet · nuScenes 30m | mAP: 62.9 -> 63.9 | Improved |
| add dynamic curve noising (ablation) | SQD-MapNet · nuScenes 30m | mAP: 61.5 -> 63.9 | Improved |
| increase matching tolerance alpha beyond 0.1 (ablation) | SQD-MapNet · nuScenes 30m | mAP: 63.5 -> 62.9 (alpha 0.2) | Hurt |
| increase noise decay scale gamma from 0.2 to 0.3 or higher (ablation) | SQD-MapNet · nuScenes 30m | mAP: 63.9 -> 62.3 (gamma 0.3) | Hurt |
| SQD-MapNet vs StreamMapNet on Argoverse2 | SQD-MapNet · Argoverse2 30m | mAP: 61.5 -> 63.3 | Improved |
Closest papers by meaning
- MAP Estimation with Denoisers: Convergence Rates and Guarantees (NeurIPS 2025 · similarity 0.84)
- MapTRv2: An End-to-End Framework for Online Vectorized HD Map Construction (IJCV 2024 · similarity 0.84)
- HIMap: HybrId Representation Learning for End-to-end Vectorized HD Map Construction (CVPR 2024 · similarity 0.82)
- You Should Learn to Stop Denoising on Point Clouds in Advance (AAAI 2025 · similarity 0.82)
- MapExpert: Online HD Map Construction with Simple and Efficient Sparse Map Element Expert (AAAI 2025 · similarity 0.81)
- Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction (CVPR 2025 · similarity 0.81)
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.
