Research map / 3D vision / Depth and stereo
Global Structure-from-Motion Meets Feedforward Reconstruction
Linfei Pan, Johannes Schönberger, 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 virtual tracks to augmented BA (ablation) | GLUEMAP A-BA · SMERF | AUC@5: 32.6 -> 54.5 | Improved |
| add virtual tracks to augmented BA (ablation) | GLUEMAP A-BA · ETH3D | AUC@1: 54.8 -> 52.6 | Hurt |
| add SIFT tracks to augmented BA (ablation) | GLUEMAP A-BA · ETH3D | AUC@1: 45.4 -> 52.6 | Improved |
| add deep tracks to augmented BA (ablation) | GLUEMAP A-BA · ETH3D | AUC@1: 46.6 -> 52.6 | Improved |
| add visual overlap ratio filtering to Doppelgangers++ filtering (ablation) | GLUEMAP motion averaging · SMERF | AUC@5: 52.7 -> 55.5 | Improved |
| add visual overlap ratio filtering to Doppelgangers++ filtering (ablation) | GLUEMAP motion averaging · ETH3D | AUC@1: 20.2 -> 20.3 | No clear effect |
| add Doppelgangers++ filtering to visual overlap filtering (ablation) | GLUEMAP motion averaging · ETH3D | AUC@1: 13.5 -> 20.3 | Improved |
| add global motion averaging to pi3 (ablation) | pi3 · ETH3D | AUC@1: 13.2 -> 20.3 | Improved |
| add augmented BA after motion averaging (ablation) | pi3 · ETH3D | AUC@1: 20.3 -> 53.1 | Improved |
| use pi3 backbone instead of VGGT (ablation) | feedforward local inference · ETH3D | AUC@1: 8.6 -> 13.2 | Improved |
| use GLUEMAP instead of pi3 | pi3 · IMC2021 (full) | AUC@5: 49.2 -> 81.3 | Improved |
| increase input views from bag 5 to full | pi3 · IMC2021 | AUC@5: 58.4 -> 49.2 | Hurt |
Closest papers by meaning
- Global Structure-from-Motion Revisited (ECCV 2024 · similarity 0.91)
- VGGSfM: Visual Geometry Grounded Deep Structure From Motion (CVPR 2024 · similarity 0.91)
- Light3R-SfM: Towards Feed-forward Structure-from-Motion (CVPR 2025 · similarity 0.90)
- MP-SfM: Monocular Surface Priors for Robust Structure-from-Motion (CVPR 2025 · similarity 0.89)
- AnyMap: Learning a General Camera Model for Structure-from-Motion with Unknown Distortion in Dynamic Scenes (CVPR 2025 · similarity 0.88)
- Structure-from-Motion with a Non-Parametric Camera Model (CVPR 2025 · similarity 0.87)
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.
