Research map / 3D vision / Human pose estimation
Uni6D: A Unified CNN Framework Without Projection Breakdown for 6D 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 UV encoding to RGB-D inputs while applying resize, crop, horizontal flip, and vertical flip (ablation) | Uni6D · YCB-Video | ADD-S: 78.87 -> 92.82 | Helped |
| Add UV encoding to RGB-D inputs while applying resize, crop, horizontal flip, and vertical flip (ablation) | Uni6D · YCB-Video | ADD(S): 65.88 -> 81.95 | Helped |
| Add plain UV coordinates to RGB-D input (ablation) | Uni6D · YCB-Video | ADD-S AUC: 90.99 -> 94.06 | Helped |
| Add inverse-projected XY coordinates to RGB-D input (ablation) | Uni6D · YCB-Video | ADD-S AUC: 90.99 -> 94.17 | Helped |
| Add positional encoding to RGB-D input (ablation) | Uni6D · YCB-Video | ADD-S AUC: 90.99 -> 93.54 | Helped |
| Add depth normals to RGB-D input (ablation) | Uni6D · YCB-Video | ADD-S AUC: 90.99 -> 93.79 | Helped |
| Add the abc head as an auxiliary training task (ablation) | Uni6D with RGB-D + Plain UV + XY + PE + NRM · YCB-Video | ADD(S) AUC: 86.93 -> 88.83 | Helped |
| Use one unified CNN with direct RT-head output and no post-processing | Uni6D · YCB-Video | FPS: 3.57 -> 25.64 | Helped |
| Use Uni6D instead of FFB6D | 6D pose estimation · LineMOD | ADD-0.1d ACC: 99.7 -> 97.03 | Hurt |
| Use Uni6D instead of FFB6D | 6D pose estimation · Occlusion LineMOD | ADD-0.1d ACC: 66.2 -> 30.71 | Hurt |
Closest papers by meaning
- Single-Stage 6D Object Pose Estimation (CVPR 2020 · similarity 0.91)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.90)
- Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · similarity 0.89)
- One2Any: One-Reference 6D Pose Estimation for Any Object (CVPR 2025 · similarity 0.88)
- Pos3R: 6D Pose Estimation for Unseen Objects Made Easy (CVPR 2025 · similarity 0.87)
- CAP-Net: A Unified Network for 6D Pose and Size Estimation of Categorical Articulated Parts from a Single RGB-D Image (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.