Research map / 3D vision / 3D generation
MV-Map: Offboard HD-Map Generation with Multi-view Consistency
Ziyang Xie, Ziqi Pang, Yu-Xiong Wang
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 |
|---|---|---|---|
| replace average fusion with uncertainty network (ablation) | MV-Map · nuScenes HD map long-range (component ablation) | mIoU (All): 37.16 -> 42.09 | Improved |
| add KL-divergence supervision to the uncertainty network (ablation) | MV-Map · nuScenes HD map long-range (component ablation) | mIoU (All): 42.09 -> 42.67 | Improved |
| add Voxel-NeRF geometry to the uncertainty network (ablation) | MV-Map · nuScenes HD map long-range (component ablation) | mIoU (All): 42.67 -> 43.22 | Improved |
| add total-variance loss to Voxel-NeRF (ablation) | MV-Map · nuScenes HD map long-range (component ablation) | mIoU (All): 43.22 -> 43.92 | Improved |
| region-centric average fusion over onboard predictions (ablation) | Onboard model · nuScenes HD map (long-range) | mIoU (All): 34.95 -> 37.16 | Improved |
| use monocular depth instead of Voxel-NeRF geometry (ablation) | MV-Map · nuScenes long-range | mIoU (All): 43.92 -> 43.69 | Hurt |
| add monocular depth to uncertainty-only fusion (ablation) | MV-Map · nuScenes long-range | mIoU (All): 42.67 -> 43.69 | Improved |
| fuse semantic maps instead of BEV features (ablation) | MV-Map · nuScenes long-range | mIoU (All): 41.77 -> 43.92 | Improved |
| increase training frames from 3 to 5 (ablation) | Uncertainty network · nuScenes long-range | mIoU (All): 43.22 -> 43.92 | Improved |
| increase training frames from 5 to 7 (ablation) | Uncertainty network · nuScenes long-range | mIoU (All): 43.92 -> 44.22 | No clear effect |
| add MV-Map offboard fusion (ablation) | VectorMapNet · nuScenes vectorized HD map | mAP (All): 42.1 -> 48.9 | Improved |
| add NMP temporal fusion | VectorMapNet · nuScenes vectorized HD map | mAP (All): 40.9 -> 44.8 | Improved |
Closest papers by meaning
- SafeMap: Robust HD Map Construction from Incomplete Observations (ICML 2025 · similarity 0.88)
- Map It Anywhere: Empowering BEV Map Prediction using Large-scale Public Datasets (NeurIPS 2024 · similarity 0.87)
- VQ-Map: Bird's-Eye-View Map Layout Estimation in Tokenized Discrete Space via Vector Quantization (NeurIPS 2024 · similarity 0.86)
- MapExpert: Online HD Map Construction with Simple and Efficient Sparse Map Element Expert (AAAI 2025 · similarity 0.85)
- Toward Real-world BEV Perception: Depth Uncertainty Estimation via Gaussian Splatting (CVPR 2025 · similarity 0.85)
- Zero-Shot Novel View and Depth Synthesis with Multi-View Geometric Diffusion (CVPR 2025 · similarity 0.85)
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.
