Research map / 3D vision / Human pose estimation
Self6D: Self-Supervised Monocular 6D Object Pose Estimation
Gu Wang, Fabian Manhardt, Jianzhun Shao, Xiangyang Ji, Nassir Navab, Federico Tombari
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 mask alignment loss L_mask (ablation) | Self6D · LineMOD | Average Recall ADD(-S) (%): 0.1 -> 58.9 | Improved |
| add geometric alignment loss L_geom (ablation) | Self6D · LineMOD | Average Recall ADD(-S) (%): 6.4 -> 58.9 | Improved |
| add MS-SSIM loss (ablation) | Self6D · LineMOD | Average Recall ADD(-S) (%): 56.3 -> 58.9 | Improved |
| add perceptual loss (ablation) | Self6D · LineMOD | Average Recall ADD(-S) (%): 58.1 -> 58.9 | Improved |
| add LAB color loss L_ab (ablation) | Self6D · LineMOD | Average Recall ADD(-S) (%): 58.5 -> 58.9 | Improved |
| apply self-supervision on unannotated real RGB-D data | Self6D-LB · LineMOD | Average Recall ADD(-S) (%): 40.1 -> 58.9 | Improved |
| apply self-supervision on unannotated real RGB-D data | Self6D-LB · LineMOD Occlusion | Average Recall ADD(-S) (%): 15.1 -> 32.1 | Improved |
| apply self-supervision on real training set | Self6D-LB · Cropped LineMOD | mean angle error (deg): 19.8 -> 15.8 | Improved |
| apply self-supervision on 10% real RGB-D images | Self6D-LB · YCB-Video (5 objects) | Average Recall ADD(-S) (%): 29.9 -> 48.6 | Improved |
| increase amount of unlabeled real data to about 40% | Self6D · HomebrewedDB | mean average recall (%): 31% -> 71% | Improved |
Later papers that compared with this method
6 later accepted papers report Self6D++ as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- Cov2Pose: Leveraging Spatial Covariance for Direct Manifold-aware 6-DoF Object Pose Estimation (CVPR 2026 · better on 1 of 1 shared result)
- Environment-Agnostic Pose: Generating Environment-independent Object Representations for 6D Pose Estimation (ICCV 2025 · better on 2 of 2 shared results)
- SMOC-Net: Leveraging Camera Pose for Self-Supervised Monocular Object Pose Estimation (CVPR 2023 · better on 4 of 5 shared results)
- TexPose: Neural Texture Learning for Self-Supervised 6D Object Pose Estimation (CVPR 2023 · better on 3 of 3 shared results)
- Keypoint-Graph-Driven Learning Framework for Object Pose Estimation (CVPR 2021 · better on 3 of 3 shared results)
- Occlusion-Aware Self-Supervised Monocular 6D Object Pose Estimation (TPAMI 2021 · better on 3 of 3 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.
- DPOD — DPOD: 6D Pose Object Detector and Refiner (ICCV 2019) · better on 2 of 2 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
- Occlusion-Aware Self-Supervised Monocular 6D Object Pose Estimation (TPAMI 2021 · similarity 0.97)
- Pseudo Flow Consistency for Self-Supervised 6D Object Pose Estimation (ICCV 2023 · similarity 0.92)
- Self-Supervised Category-Level 6D Object Pose Estimation with Deep Implicit Shape Representation (AAAI 2022 · similarity 0.91)
- SO-Pose: Exploiting Self-Occlusion for Direct 6D Pose Estimation (ICCV 2021 · similarity 0.90)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.89)
- One2Any: One-Reference 6D Pose Estimation for Any Object (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.
