Research map / Graph learning / Graph algorithms and clustering
On the Recovery of Cameras from Fundamental Matrices
Rakshith Madhavan, Federica Arrigoni
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 GPSFM with ANGLE-IT-IRLS for camera recovery | GPSFM · 25 real projective SfM sequences | reprojection error: better in 20 out of 25 sequences | Improved |
| replace GPSFM-SYNCH with ANGLE-IT-IRLS for camera recovery | GPSFM-SYNCH · 25 real projective SfM sequences | reprojection error: better in 21 out of 25 cases | Improved |
| replace GPSFM-SYNCH with ANGLE-IT-IRLS | GPSFM-SYNCH · Toronto University | reprojection error (px): 18.28 -> 64.10 | Hurt |
| replace GPSFM-SYNCH with ANGLE-IT-IRLS | GPSFM-SYNCH · Alcatraz Water Tower | reprojection error (px): 19.76 -> 36.54 | Hurt |
| replace GPSFM with ANGLE-IT-IRLS | GPSFM · Nijo | reprojection error (px): 29.85 -> 17.38 | Improved |
| replace GPSFM-SYNCH with ANGLE-IT-IRLS | GPSFM-SYNCH · De Guerre | reprojection error (px): 1.33 -> 1.82 | Hurt |
| use ANGLE-IT-IRLS on pruned graph with a camera not in any triplet | GPSFM · Sphinx (pruned graph) | reprojection error (px): 8.16 -> 4.71 | Improved |
| use ANGLE-IT-IRLS on pruned graph with a camera not in any triplet | GPSFM-SYNCH · Sphinx (pruned graph) | reprojection error (px): 7.10 -> 4.71 | Improved |
| add IRLS robust reweighting (ablation) | LS-SVD · synthetic, increasing outlier density | camera error (degrees) | Improved |
| add IRLS robust reweighting (ablation) | ANGLE-IT · synthetic, increasing outlier density | camera error (degrees) | Improved |
| use unsquared angular distance (ANGLE-IT) instead of least squares (LS-SVD) (ablation) | alternating optimization · synthetic, outliers up to 40% without IRLS | camera error (degrees) | Improved |
| use angular distance (ANGLE-IT) instead of least squares (LS-SVD) (ablation) | alternating optimization · synthetic, increasing noise (n=25, rho=0.4) | camera error (degrees) | No clear effect |
Closest papers by meaning
- On the Solvability of Viewing Graphs (ECCV 2018 · similarity 0.92)
- A Direct Approach to Viewing Graph Solvability (ECCV 2024 · similarity 0.90)
- Three-view Focal Length Recovery From Homographies (CVPR 2025 · similarity 0.84)
- Leveraging Camera Triplets for Efficient and Accurate Structure-from-Motion (CVPR 2024 · similarity 0.84)
- Graph-Guided Scene Reconstruction from Images with 3D Gaussian Splatting (ICLR 2025 · similarity 0.83)
- Learning to Filter Outlier Edges in Global SfM (CVPR 2025 · similarity 0.83)
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.
