Research map / Learning theory and algorithms / Combinatorial and search optimization
Closed-Form Optimal Two-View Triangulation Based on Angular Errors
Seong Hun Lee, Javier Civera
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 |
|---|---|---|---|
| use angular-error L1 closed form instead of image-error L1 (Hartley-Sturm) | L1 triangulation · Triangulation speed | relative speed: 0.0016 -> 0.71 | Improved |
| use angular-error Linf closed form instead of Nister's image-error Linf | Linf triangulation · Triangulation speed | relative speed: 0.0064 -> 0.33 | Improved |
| use angular sine-relaxed L2 instead of Hartley-Sturm L2 | L2 triangulation · Triangulation speed | relative speed: 0.0022 -> 0.016 | Improved |
| use angular L2 instead of Lindstrom L2 with five iterations | L2 triangulation · Triangulation speed | relative speed: 0.013 -> 0.016 | Improved |
| angular L1 closed form vs Hartley-Sturm L1 on the image L1 criterion | L1 triangulation · Image L1 error d0+d1 | % matches with lowest error: 70.84% vs 29.16 % | Mixed |
| minimize sin^2 instead of theta^2 (small-angle relaxation) | angular L2 triangulation · Angular L2 criterion | % matches with lowest theta0^2+theta1^2: 99.9999 % | No clear effect |
| discard points with lowest 5% parallax (ablation) | all triangulation methods · Synthetic datasets (orbital, lateral, forward) | 3D reconstruction error histogram | Improved |
| compare methods after discarding low-parallax points | all triangulation methods · Synthetic datasets | 3D accuracy | No clear effect |
| compare methods on real datasets | all triangulation methods · Real datasets (Dinosaur, Model House, Corridor, Notre Dame, Fountain) | qualitative reconstruction | No clear effect |
Closest papers by meaning
- Bound and Conquer: Improving Triangulation by Enforcing Consistency (TPAMI 2019 · similarity 0.88)
- Radial Distortion Triangulation (CVPR 2019 · similarity 0.87)
- Semidefinite Relaxations for Robust Multiview Triangulation (CVPR 2023 · similarity 0.85)
- Tangent Sampson Error: Fast Approximate Two-view Reprojection Error for Central Camera Models (ICCV 2023 · similarity 0.85)
- Practical Solutions to the Relative Pose of Three Calibrated Cameras (CVPR 2025 · similarity 0.83)
- Minimal Perspective Autocalibration (CVPR 2024 · 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.
