Research map / 3D vision / Human pose estimation
Focal Length and Object Pose Estimation via Render and Compare
Georgy Ponimatkin, Yann Labbé, Bryan Russell, Mathieu Aubry, Josef Sivic
Check 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 |
|---|---|---|---|
| Add disentangled reprojection loss to Huber focal-length regression (ablation) | FocalPose · Pix3D sofa | MedErr_R: 3.28 -> 2.98 | Improved |
| Add disentangled reprojection loss to Huber focal-length regression (ablation) | FocalPose · Pix3D sofa | MedErr_t: 1.42 -> 1.29 | Improved |
| Add disentangled reprojection loss to Huber focal-length regression (ablation) | FocalPose · Pix3D sofa | MedErr_f: 1.45 -> 1.36 | Improved |
| Add reprojection-based supervision to Huber focal-length regression (ablation) | FocalPose · Pix3D sofa | MedErr_R: 6.61 -> 2.98 | Improved |
| Add reprojection-based supervision to Huber focal-length regression (ablation) | FocalPose · Pix3D sofa | MedErr_f: 4.17 -> 1.36 | Improved |
| Use a mix of mostly synthetic data and a small amount of real data rather than synthetic-only or real-only data (ablation) | FocalPose coarse model · Pix3D sofa | MedErr_R: 5.44 / 4.13 -> 2.98 | Improved |
| Use a mix of mostly synthetic data and a small amount of real data rather than synthetic-only or real-only data (ablation) | FocalPose coarse model · Pix3D sofa | MedErr_t: 2.18 / 1.92 -> 1.29 | Improved |
| Use a mix of mostly synthetic data and a small amount of real data rather than synthetic-only or real-only data (ablation) | FocalPose coarse model · Pix3D sofa | MedErr_f: 2.04 / 1.91 -> 1.36 | Improved |
| Increase the real-data fraction in mixed training batches from 0.5% to 5% (ablation) | FocalPose coarse model · Pix3D sofa | MedErr_R: 2.98 -> 3.08 | Hurt |
| Use render-and-compare with jointly estimated focal length and pose | FocalPose · Pix3D, CompCars, Stanford Cars | MedErr_f: Pix3D 1.72 -> 1.53; CompCars 2.95 -> 2.65; Stanford Cars 2.01 -> 1.09 | Improved |
| Use render-and-compare with jointly estimated focal length and pose | FocalPose · Pix3D, CompCars, Stanford Cars | MedErr_t: Pix3D 1.85 -> 1.49; CompCars 2.55 -> 2.35; Stanford Cars 1.92 -> 1.00 | Improved |
| Run more refinement iterations at inference (ablation) | FocalPose · Pix3D and car datasets | MedErr_t | Improved |
Closest papers by meaning
- RefPose: Leveraging Reference Geometric Correspondences for Accurate 6D Pose Estimation of Unseen Objects (CVPR 2025 · similarity 0.88)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.87)
- Pos3R: 6D Pose Estimation for Unseen Objects Made Easy (CVPR 2025 · similarity 0.87)
- Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · similarity 0.87)
- FAR: Flexible Accurate and Robust 6DoF Relative Camera Pose Estimation (CVPR 2024 · similarity 0.87)
- One2Any: One-Reference 6D Pose Estimation for Any Object (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.
