Research map / 3D vision / Human pose estimation
BoxDreamer: Dreaming Box Corners for Generalizable Object Pose Estimation
Yuanhong Yu, Xingyi He, Chen Zhao, Junhao Yu, Jiaqi Yang, Ruizhen Hu, Yujun Shen, Xing Zhu, Xiaowei Zhou, Sida Peng
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 real-data (OnePose) training (ablation) | BoxDreamer · LINEMOD (5 refs) | ADD(s)-0.1d: 40.6 -> 51.3 | Improved |
| add real-data (OnePose) training (ablation) | BoxDreamer · LINEMOD (5 refs) | Proj2D@5px: 25.4 -> 43.8 | Improved |
| use DUSt3R box instead of ground-truth box (ablation) | BoxDreamer · LINEMOD (5 refs) | ADD(s)-0.1d: 49.3 -> 51.3 | No clear effect |
| use DUSt3R box instead of ground-truth box (ablation) | BoxDreamer · LINEMOD (5 refs) | Proj2D@5px: 51.8 -> 43.8 | Hurt |
| use OnePose++ box instead of DUSt3R box (ablation) | BoxDreamer · LINEMOD (5 refs) | Proj2D@5px: 43.8 -> 52.7 | Improved |
| increase references from 5 to 32 (ablation) | BoxDreamer · LINEMOD subset | ADD(s)-0.1d: 53.1 -> 69.2 | Improved |
| BoxDreamer vs GS-Pose at 5 references | BoxDreamer · LINEMOD subset (5 refs) | ADD(s)-0.1d: 25.6 -> 53.1 | Improved |
| BoxDreamer vs GS-Pose at 32 references | BoxDreamer · LINEMOD subset (32 refs) | ADD(s)-0.1d: 74.5 -> 69.2 | Hurt |
| BoxDreamer vs OnePose++ with 10 references | BoxDreamer · OnePose-LowTexture | ADD-0.1d: 58.0 -> 86.6 | Improved |
| use denser 200-reference database for baselines | Gen6D · YCB-Video most-overlapping database | ADD: 14.6 -> 22.9 | No clear effect |
| use occlusion-minimizing reference database instead of most-overlapping | BoxDreamer · YCB-Video | ADD: 35.4 -> 37.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- OnePose++ — OnePose: One-Shot Object Pose Estimation Without CAD Models (CVPR 2022) · better on 5 of 5 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
- Pose without Guesses: Generalizable Object Pose Estimation From a Single Reference (IJCV 2026 · similarity 0.87)
- 3D-Aware Hypothesis & Verification for Generalizable Relative Object Pose Estimation (ICLR 2024 · similarity 0.86)
- UNOPose: Unseen Object Pose Estimation with an Unposed RGB-D Reference Image (CVPR 2025 · similarity 0.85)
- Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis (NeurIPS 2024 · similarity 0.85)
- Co-op: Correspondence-based Novel Object Pose Estimation (CVPR 2025 · similarity 0.85)
- GigaPose: Fast and Robust Novel Object Pose Estimation via One Correspondence (CVPR 2024 · similarity 0.84)
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.
