Research map / 3D vision / Depth and stereo
RelPose: Predicting Probabilistic Relative Rotation for Single Objects in the Wild
Jason Y. Zhang, Deva Ramanan, Shubham Tulsiani
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 |
|---|---|---|---|
| add maximum spanning tree linking instead of sequential chain (ablation) | Ours · CO3D seen categories (random sampling) | Acc @ 30 deg: 0.39 -> 0.43 (N=20) | Improved |
| add coordinate ascent joint refinement on top of MST init (ablation) | Ours (MST) · CO3D seen categories (random sampling) | Acc @ 30 deg: 0.43 -> 0.59 (N=20) | Improved |
| use coordinate ascent instead of gradient ascent (ablation) | Ours · CO3D seen categories (random sampling) | Acc @ 30 deg: 0.47 -> 0.59 (N=20) | Improved |
| add gradient ascent from MST init (ablation) | Ours (MST) · CO3D seen categories (random sampling) | Acc @ 30 deg: 0.52 -> 0.52 (N=3) | No clear effect |
| increase number of images with MST only | Ours (MST) · CO3D seen categories (random sampling) | Acc @ 30 deg: 0.52 -> 0.43 (N=3 to N=20) | Hurt |
| increase number of images with coordinate ascent | Ours (Coord. Asc.) · CO3D seen categories (random sampling) | Acc @ 30 deg: 0.59 -> 0.59 (N=3 to N=20) | No clear effect |
| evaluate on unseen instead of seen categories | Ours (Coord. Asc.) · CO3D per-category mean, random sampling | Acc @ 30 deg (%): 61 -> 49 (N=3) | Hurt |
| use energy-based probabilistic prediction instead of direct regression of relative rotation (ablation) | Ours · CO3D seen categories | Acc @ 15/30 deg | Improved |
Closest papers by meaning
- Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis (NeurIPS 2024 · similarity 0.88)
- Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localization (CVPR 2025 · similarity 0.87)
- SuperPrimitive: Scene Reconstruction at a Primitive Level (CVPR 2024 · similarity 0.87)
- LEAP: Liberate Sparse-View 3D Modeling from Camera Poses (ICLR 2024 · similarity 0.87)
- 3D-Aware Hypothesis & Verification for Generalizable Relative Object Pose Estimation (ICLR 2024 · similarity 0.87)
- From Correspondences to Pose: Non-minimal Certifiably Optimal Relative Pose without Disambiguation (CVPR 2024 · similarity 0.86)
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.
