Research map / 3D vision / Human pose estimation
Object Level Depth Reconstruction for Category Level 6D Object Pose Estimation from Monocular RGB Image
Zhaoxin Fan, Zhenbo Song, Jian Xu, Zhicheng Wang, Kejian Wu, Hongyan Liu, Jun He
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 SDDR and NGPH on top of vanilla SPD (full model) (ablation) | OLD-Net · CAMERA25 | 10 deg 10cm: 4.4 -> 23.4 | Improved |
| add decoupled depth translation design (ablation) | OLD-Net · CAMERA25 | IoU25: 72.7 -> 74.3 | Improved |
| add decoupled depth translation design (ablation) | OLD-Net · CAMERA25 | IoU75: 5.5 -> 5.4 | No clear effect |
| use SPD shape points instead of MLP (ablation) | OLD-Net · CAMERA25 | 10 deg 10cm: 20.7 -> 23.4 | Improved |
| add NGPH (ablation) | OLD-Net · CAMERA25 | 10 deg: 61.6 -> 74.0 | Improved |
| add NGPH (ablation) | OLD-Net · CAMERA25 | IoU50: 29.4 -> 32.1 | Improved |
| add adversarial training (ablation) | OLD-Net · CAMERA25 | 10 deg 10cm: 22.1 -> 23.4 | Improved |
| add adversarial training (ablation) | OLD-Net · CAMERA25 | 10 deg: 75.1 -> 74.0 | Hurt |
| predict object-level depth instead of scene-level depth (shared encoder) | OLD-Net · CAMERA25 | 10 deg 10cm: 1.6 -> 23.4 | Improved |
| predict object-level depth instead of scene-level depth (independent) | OLD-Net · CAMERA25 | IoU25: 50.4 -> 74.3 | Improved |
| compare to Lee et al. mesh-render depth | OLD-Net · CAMERA25 | 10 deg 10cm: 19.2 -> 23.4 | Improved |
| compare to Lee et al. | OLD-Net · REAL275 | IoU25: 62.0 -> 68.7 | Improved |
Closest papers by meaning
- Boosting 6D Object Pose Estimation via Monocular Depth Cues (ECCV 2026 · similarity 0.92)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.89)
- Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation (CVPR 2019 · similarity 0.89)
- Leveraging Global Stereo Consistency for Category-Level Shape and 6D Pose Estimation from Stereo Images (CVPR 2025 · similarity 0.89)
- GIVEPose: Gradual Intra-class Variation Elimination for RGB-based Category-Level Object Pose Estimation (CVPR 2025 · similarity 0.86)
- Universal Features Guided Zero-Shot Category-Level Object Pose Estimation (AAAI 2025 · 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.
