Research map / 3D vision / Human pose estimation
OVE6D: Object Viewpoint Encoding for Depth-Based 6D Object Pose Estimation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 ICP refinement to OVE6D pose proposals before selection (ablation) | OVE6D · LINEMOD | ADD(-S) recall (%): 96.4 -> 98.7 | Helped |
| Add ICP refinement to OVE6D pose proposals before selection (ablation) | OVE6D · T-LESS | VSD recall (%): 85.1 -> 91.0 | Helped |
| Use ground-truth masks instead of Mask R-CNN masks (ablation) | OVE6D · T-LESS | VSD recall (%): 74.8 -> 91.0 | Helped |
| Increase viewpoint codebook samples from 1k to 16k (ablation) | OVE6D · LINEMOD, K=1 and P=1 | ADD(-S) recall (%): 74.1 -> 76.6 | Helped |
| Retrieve more viewpoint candidates (K=1 to K=50) (ablation) | OVE6D · LINEMOD, N=4k and P=1 | ADD(-S) recall (%): 75.7 -> 81.6 | Helped |
| Retrieve 100 rather than 50 viewpoint candidates (ablation) | OVE6D · LINEMOD, N=4k and P=1 | ADD(-S) recall (%): 81.6 -> 81.5 | No clear effect |
| Increase orientation proposals from P=1 to P=20 (ablation) | OVE6D · LINEMOD, N=4k and K=50 | ADD(-S) recall (%): 81.6 -> 87.2 | Helped |
| Build the viewpoint codebook from real annotated images instead of the object mesh (ablation) | OVE6D · LINEMOD, ICP before selection | ADD(-S) recall (%): 98.7 -> 99.4 | Helped |
| Use the proposed location refinement rather than the initial location estimate (ablation) | OVE6D · LINEMOD | precision at 10mm error tolerance: 13% -> 60% | Helped |
| Add ICP refinement after the proposed location refinement (ablation) | OVE6D · LINEMOD | precision at 10mm error tolerance: 60% -> 81% | Helped |
| Use the ground-truth viewpoint for in-plane orientation estimation (ablation) | OVE6D · LINEMOD | in-plane orientation precision at 10° error threshold: up to 73% -> over 90% | Helped |
Closest papers by meaning
- One2Any: One-Reference 6D Pose Estimation for Any Object (CVPR 2025 · similarity 0.89)
- Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · similarity 0.89)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.89)
- FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · similarity 0.88)
- Open-Vocabulary Object 6D Pose Estimation (CVPR 2024 · similarity 0.88)
- Pos3R: 6D Pose Estimation for Unseen Objects Made Easy (CVPR 2025 · similarity 0.88)
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.