Research map / 3D vision / Depth and stereo
Uncalibrated Structure from Motion on a Sphere
Jonathan Ventura, Viktor Larsson, Fredrik Kahl
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 |
|---|---|---|---|
| skip the final general bundle adjustment so cameras stay on a perfect sphere (ablation) | Ours · Deep View Video | AUC@30: 83.87 -> 91.99 | Hurt |
| skip the final general bundle adjustment (ablation) | Ours · Deep View Video | RTA@5: 74.53 -> 93.63 | Hurt |
| skip the final general bundle adjustment (ablation) | Ours · Deep View Video | Recall: 87.74 -> 93.12 | Hurt |
| skip the final general bundle adjustment (ablation) | Ours · Deep View Video | RRA@5: 99.71 -> 99.56 | No clear effect |
| use three-point spherical solver instead of seven/eight-point fundamental matrix solvers | Ours · synthetic spherical motion | fundamental matrix error | Improved |
| use spherical uncalibrated SfM instead of COLMAP with unknown focal length | Ours · PhoneSweep (combined) | AUC@30: 67.04 -> 90.87 | Improved |
| use spherical uncalibrated SfM instead of GLOMAP | Ours · PhoneSweep (combined) | AUC@30: 55.79 -> 90.87 | Improved |
| compare to COLMAP given the ground-truth focal length | Ours · PhoneSweep (combined) | AUC@30: 83.16 -> 90.87 | Improved |
| use spherical uncalibrated SfM instead of COLMAP on hemispherical rig | Ours · Deep View Video | AUC@30: 36.44 -> 83.87 | Improved |
| use spherical SfM instead of COLMAP for focal length estimation | Ours · PhoneSweep (combined) | absolute focal error: 2.95 -> 0.58 | Improved |
| use spherical SfM instead of GLOMAP for runtime | Ours · PhoneSweep | time per image: 2.56 -> 3.97 | Hurt |
| use spherical SfM instead of GLOMAP for runtime | Ours · Deep View Video | time per image: 1.73 -> 1.13 | Improved |
Closest papers by meaning
- Structure from Motion on a Sphere (ECCV 2016 · similarity 0.91)
- Calibration-free Structure-from-Motion with Calibrated Radial Trifocal Tensors (ECCV 2020 · similarity 0.85)
- Planar Structure-from-Motion with Affine Camera Models: Closed-Form Solutions, Ambiguities and Degeneracy Analysis (TPAMI 2016 · similarity 0.85)
- Structure-from-motion in micro-image domain for uncalibrated plenoptic 2.0 cameras (IJCV 2026 · similarity 0.85)
- Non-Rigid Structure-from-Motion: Temporally-Smooth Procrustean Alignment and Spatially-Variant Deformation Modeling (CVPR 2024 · similarity 0.84)
- Structure-from-Motion with a Non-Parametric Camera Model (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.
