Research map / Detection and segmentation / 3D detection and driving
Map-Det3D: Metric Feed-Forward 3D Reconstruction Prior for Multi-view 3D Object Detection from Streaming Inputs
Yung-Hsu Yang, Luigi Piccinelli, Samuel Rota Bulò, Sunghwan Hong, Denys Rozumnyi, Johannes Schönberger, Zuria Bauer, Hermann Blum, Peter Kontschieder, Marc Pollefeys
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 multi-view input with frozen backbone (ablation) | Map-Det3D · CA-1M val | AP15: 11.7 -> 11.8 | No clear effect |
| Unfreeze scale head only (ablation) | Map-Det3D · CA-1M val | AP15: 11.8 -> 11.9 | No clear effect |
| Unfreeze multi-view transformer (and scale head) (ablation) | Map-Det3D · CA-1M val | AP15: 11.9 -> 14.5 | Improved |
| Add multi-view input after unfreezing (ablation) | Map-Det3D · CA-1M val | AP15: 14.5 -> 17.2 | Improved |
| Condition on camera intrinsics (ablation) | Map-Det3D · CA-1M val | AP15: 11.8 -> 12.6 | Improved |
| Add camera poses to conditioning (ablation) | Map-Det3D · CA-1M val | AP15: 12.6 -> 17.3 | Improved |
| Enable all components (full model) (ablation) | Map-Det3D · CA-1M val | AP15: 11.7 -> 21.2 | Improved |
| Add direct up-to-scale 3D box head instead of 2D-to-3D lifting (ablation) | Map-Det3D · CA-1M val | AP15: 15.6 -> 17.3 | Improved |
| Add direct up-to-scale 3D box head instead of 2D-to-3D lifting (ablation) | Map-Det3D · ScanNet (zero-shot) | AP15: 10.8 -> 12.6 | Improved |
| Increase window from T=1 to T=5 (ablation) | Map-Det3D · CA-1M val | AP15: 14.0 -> 21.2 | Improved |
| Test with T=7 instead of trained T=5 (ablation) | Map-Det3D · CA-1M val | AP15: 21.2 -> 21.1 | No clear effect |
| Offline few-view setting (strides=5) instead of online sliding window (ablation) | Map-Det3D · CA-1M val | AP15: 21.2 -> 22.2 | 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.
- 3D-MOOD — 3D-MOOD: Lifting 2D to 3D for Monocular Open-Set Object Detection (ICCV 2025) · better on 2 of 2 shared results
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
- Detect Anything 3D in the Wild (ICCV 2025 · similarity 0.91)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 2025 · similarity 0.90)
- Training an Open-Vocabulary Monocular 3D Detection Model without 3D Data (NeurIPS 2024 · similarity 0.89)
- Rooms from Motion: Un-posed Indoor 3D Object Detection as Localization and Mapping (NeurIPS 2025 · similarity 0.88)
- MonoDiff: Monocular 3D Object Detection and Pose Estimation with Diffusion Models (CVPR 2024 · similarity 0.88)
- MonoPlace3D: Learning 3D-Aware Object Placement for 3D Monocular Detection (CVPR 2025 · 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.
