Research map / 3D vision / Human pose estimation
Boosting 6D Object Pose Estimation via Monocular Depth Cues
Fengda Hao, Rui Song, Qingyuan Wang, Jiaojiao Li, Zhiyong Hu, David Ferstl, Yinlin Hu
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 SRM (scale refinement) (ablation) | MDC-Net · LM-O | AR: 59.9 -> 63.0 | Improved |
| add IGDR (inlier-guided depth refinement) (ablation) | MDC-Net · LM-O | AR: 59.9 -> 64.5 | Improved |
| add both SRM and IGDR (ablation) | MDC-Net · LM-O | AR: 59.9 -> 66.8 | Improved |
| add IGDR on top of SRM (ablation) | MDC-Net · LM-O | AR: 63 -> 66.8 | Improved |
| add SRM on top of IGDR (ablation) | MDC-Net · LM-O | AR: 64.5 -> 66.8 | Improved |
| add both SRM and IGDR (ablation) | MDC-Net · YCB-V | AR: 66.3 -> 74.1 | Improved |
| add IGDR to SRM (ablation) | MDC-Net · YCB-V | VSD: 64.4 -> 61.6 | Hurt |
| use MoGe-2 metric depth instead of ZoeDepth relative depth (ablation) | MDC-Net · LM-O (Table 3 column) | AR: 68.7 -> 74.1 | Improved |
| use ground-truth depth instead of MoGe-2 depth (ablation) | MDC-Net · LM-O (Table 3 column) | AR: 74.1 -> 85.5 | Improved |
| replace FoundPose refinement stage with MDC-Net | FoundPose · LM-O | AR: 61 -> 64.2 | Improved |
| replace SCFlow refinement stage with MDC-Net | SCFlow · LM-O | AR: 64.9 -> 66.8 | Improved |
| replace GenFlow refinement stage with MDC-Net | GenFlow · YCB-V | AR: 70.9 -> 73.3 | 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.
- GigaPose — GigaPose: Fast and Robust Novel Object Pose Estimation via One Correspondence (CVPR 2024) · 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
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.92)
- Leveraging Global Stereo Consistency for Category-Level Shape and 6D Pose Estimation from Stereo Images (CVPR 2025 · similarity 0.89)
- Relative Pose Estimation through Affine Corrections of Monocular Depth Priors (CVPR 2025 · similarity 0.89)
- RefPose: Leveraging Reference Geometric Correspondences for Accurate 6D Pose Estimation of Unseen Objects (CVPR 2025 · similarity 0.88)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.87)
- Any6D: Model-free 6D Pose Estimation of Novel Objects (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.
