Research map / 3D vision / Human pose estimation
Minimal Cases for Computing the Generalized Relative Pose Using Affine Correspondences
Banglei Guan, Ji Zhao, Daniel Barath, Friedrich Fraundorfer
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 2AC vertical (known vertical direction) instead of 2AC plane | affine-correspondence solver · KITTI sequence 0 | rotation error (deg): 0.280 -> 0.031 | Improved |
| use hallucinated point correspondences converted from ACs instead of the AC image point pairs (ablation) | PC-based solvers · synthetic random motion, known vertical | pose error | Hurt |
| reduce the distribution area w of hallucinated PCs (ablation) | PC-based solvers with hallucinated PCs · synthetic random motion, known vertical | pose error: w = 1, 5, 10 pixels | Improved |
| increase the affine support square side from 20 to 40 pixels (ablation) | 1AC plane · synthetic planar motion | translation direction error: 20 -> 40 pixels | Improved |
| add a second AC to the planar solver (2AC plane instead of 1AC plane) (ablation) | planar motion solver · synthetic planar motion | translation error | Improved |
| increase planar motion noise (ablation) | 2AC plane · synthetic planar motion noise | rotation error: less than 0.2 deg noise: better than comparisons | Hurt |
| use an AC-based minimal solver (1 or 2 points) instead of larger point solvers (ablation) | RANSAC · outlier ratio 50% | RANSAC iterations: 7, 16, 71, 292, 1177 and 603607 iterations | Improved |
| apply planar-motion assumption on non-planar KITTI data | 2AC plane · KITTI | translation error | No clear effect |
Closest papers by meaning
- Minimal Solvers for Relative Pose Estimation of Multi-Camera Systems using Affine Correspondences (IJCV 2022 · similarity 0.98)
- A Complete Solution to Generalized Relative Pose Estimation From Affine Correspondences (TPAMI 2026 · similarity 0.95)
- Affine Correspondences Between Multi-Camera Systems for Relative Pose Estimation (TPAMI 2025 · similarity 0.94)
- From Correspondences to Pose: Non-minimal Certifiably Optimal Relative Pose without Disambiguation (CVPR 2024 · similarity 0.90)
- Practical Solutions to the Relative Pose of Three Calibrated Cameras (CVPR 2025 · similarity 0.88)
- Efficient Solution of Point-Line Absolute Pose (CVPR 2024 · 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.
