Research map / 3D vision / Human pose estimation
P1AC: Revisiting Absolute Pose From a Single Affine Correspondence
Jonathan Ventura, Zuzana Kukelova, Torsten Sattler, Dániel Baráth
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 P3P with P1AC as minimal solver in GC-RANSAC | GC-RANSAC localization · Cambridge Landmarks avg | recall 0.1m/1deg (%): 54.62 -> 65.08 | Improved |
| replace P3P with P1AC as minimal solver in GC-RANSAC | GC-RANSAC localization · Aachen Day-Night Night | recall 5m/10deg (%): 74.3 -> 82.2 | Improved |
| replace P3P with P1AC on Aachen Day sequence | GC-RANSAC localization · Aachen Day-Night Day | recall 0.25m/2deg (%): 62.0 -> 62.0 | No clear effect |
| use P1AC instead of DPR and IPPE (no virtual warping) | single-AC solvers · Cambridge Landmarks avg | recall 0.1m/1deg (%): 63.48 (DPR) -> 65.08 | Improved |
| increase perspective skew of affine correspondence (ablation) | DPR solver · synthetic 1M instances | rotation accuracy | Hurt |
| use vanilla RANSAC with P1AC instead of P3P at realistic noise | synthetic robust estimation · synthetic outlier ratios 0-0.9 | average rotation error (deg): 0.056 -> 0.450 | Hurt |
| add local optimization (LO-RANSAC) with P1AC (ablation) | P1AC · synthetic robust estimation | average rotation error (deg): 0.450 -> 0.085 | Improved |
| drop SNN ratio filtering for P1AC on Cambridge (ablation) | P1AC · Cambridge Landmarks | accuracy and inlier count | Improved |
| add affine and normal noise to P1AC input (ablation) | P1AC solver · synthetic noise | pose error | Hurt |
| P1AC runtime in GC-RANSAC vs P3P | GC-RANSAC · Cambridge Landmarks | time (s): Great Court 0.14 -> 0.26 | Hurt |
Later papers that compared with this method
1 later accepted paper reports P1AC as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Absolute Pose from One or Two Scaled and Oriented Features (CVPR 2024 · better on 7 of 7 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- From Correspondences to Pose: Non-minimal Certifiably Optimal Relative Pose without Disambiguation (CVPR 2024 · similarity 0.89)
- Efficient Solution of Point-Line Absolute Pose (CVPR 2024 · similarity 0.89)
- Relative Pose Estimation through Affine Corrections of Monocular Depth Priors (CVPR 2025 · similarity 0.88)
- Practical Solutions to the Relative Pose of Three Calibrated Cameras (CVPR 2025 · similarity 0.87)
- Absolute Pose from One or Two Scaled and Oriented Features (CVPR 2024 · similarity 0.87)
- Learning Affine Correspondences by Integrating Geometric Constraints (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.
