Research map / 3D vision / Human pose estimation
Gen6D: Generalizable Model-Free 6-DoF Object Pose Estimation from RGB Images
Yuan Liu, Yilin Wen, Sida Peng, Cheng Lin, Xiaoxiao Long, Taku Komura, Wenping Wang
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 the volume-based pose refiner (ablation) | Gen6D · GenMOP | ADD-0.1d: 17.90 -> 50.39 | Improved |
| add the volume-based pose refiner (ablation) | Gen6D · GenMOP | Prj-5: 36.73 -> 82.64 | Improved |
| add global normalization and reference view transformer to the selector (ablation) | Viewpoint selector · GenMOP | ADD-0.1d: 14.88 -> 17.90 | Improved |
| add global normalization and reference view transformer to the selector (ablation) | Viewpoint selector · GenMOP | Prj-5: 30.28 -> 36.73 | Improved |
| replace DeepIM refiner with volume refiner (ablation) | Gen6D · GenMOP | Prj-5: 42.16 -> 77.54 | Improved |
| replace pretrained DeepIM with volume refiner (ablation) | Gen6D · GenMOP | ADD-0.1d: 19.53 -> 50.39 | Improved |
| add real-data (LINEMOD) training to synthetic+GenMOP training (ablation) | Gen6D · GenMOP | ADD-0.1d: 39.94 -> 50.39 | Improved |
| add GenMOP real data to synthetic training (ablation) | Gen6D · GenMOP | ADD-0.1d: 33.70 -> 39.94 | Improved |
| use 3 refinement iterations instead of 2 (ablation) | Gen6D · GenMOP | ADD-0.1d: 53.32 -> 50.39 | No clear effect |
| use 1 refinement iteration instead of none (ablation) | Gen6D · GenMOP | ADD-0.1d: 17.90 -> 38.59 | Improved |
| reduce reference images from all to 8 (ablation) | Gen6D · GenMOP | Prj-5: 82.64 -> 37.33 | Hurt |
| reduce reference images from all to 64 (ablation) | Gen6D · GenMOP | Prj-5: 82.64 -> 82.30 | No clear effect |
Later papers that compared with this method
3 later accepted papers report Gen6D as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- PoseGaussian: 6D Pose Estimation for Unseen Objects via Sparse-View Object-Level 3D Gaussian Splatting (CVPR 2026 · better on 1 of 1 shared result)
- Prior-free 3D Object Tracking (CVPR 2025 · better on 1 of 1 shared result)
- VoxDet: Voxel Learning for Novel Instance Detection (NeurIPS 2023 · better on 2 of 2 shared results)
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.
- PVNet — PVNet: Pixel-Wise Voting Network for 6DoF Pose Estimation (CVPR 2019) · better on 3 of 3 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
- Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · similarity 0.89)
- GenFlow: Generalizable Recurrent Flow for 6D Pose Refinement of Novel Objects (CVPR 2024 · similarity 0.88)
- Pos3R: 6D Pose Estimation for Unseen Objects Made Easy (CVPR 2025 · similarity 0.88)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.87)
- One2Any: One-Reference 6D Pose Estimation for Any Object (CVPR 2025 · similarity 0.87)
- FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (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.
