Research map / 3D vision / Human pose estimation
Perspective Flow Aggregation for Data-Limited 6D Object Pose Estimation
Yinlin Hu, Pascal Fua, Mathieu Salzmann
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 noise augmentation (NS) (ablation) | WDR-Pose initialization · LINEMOD | ADD-0.1d: 48.9 -> 56.6 | Improved |
| add color augmentation (HSV) (ablation) | WDR-Pose initialization · LINEMOD | ADD-0.1d: 48.9 -> 58.1 | Improved |
| add NS to HSV augmentation (ablation) | WDR-Pose initialization · YCB-V | ADD-0.1d: 7.4 -> 27.5 | Improved |
| add geometric shift-scale-rotate augmentation (SSR) (ablation) | WDR-Pose initialization · LINEMOD | ADD-0.1d: 48.9 -> 39.0 | Hurt |
| add geometric shift-scale-rotate augmentation (SSR) (ablation) | WDR-Pose initialization · Occluded-LINEMOD | ADD-0.1d: 28.6 -> 20.9 | Hurt |
| add flow-based refinement with one exemplar to the initial pose (ablation) | Ours · LINEMOD | ADD-0.1d: 60.2 -> 82.0 | Improved |
| increase retrieved exemplars from 1 to 4 (ablation) | Ours · LINEMOD | ADD-0.1d: 82.0 -> 84.5 | Improved |
| increase retrieved exemplars from 4 to 8 (ablation) | Ours · LINEMOD | ADD-0.1d: 84.5 -> 84.9 | No clear effect |
| increase exemplar set size from 2.5K to 10K (ablation) | Ours · LINEMOD | ADD-0.1d: 81.9 -> 84.5 | Improved |
| increase exemplar set size from 10K to 40K (ablation) | Ours · LINEMOD | ADD-0.1d: 84.5 -> 84.8 | No clear effect |
| use offline exemplars instead of online rendering (ablation) | Ours · LINEMOD | ADD-0.1d: 84.9 -> 84.5 | No clear effect |
| add 20 real images to synthetic training | Ours · Occluded-LINEMOD | ADD-0.1d: 48.2 -> 64.1 | Improved |
Closest papers by meaning
- Pseudo Flow Consistency for Self-Supervised 6D Object Pose Estimation (ICCV 2023 · similarity 0.90)
- Single-Stage 6D Object Pose Estimation (CVPR 2020 · similarity 0.88)
- Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · similarity 0.87)
- FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · similarity 0.86)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.86)
- One2Any: One-Reference 6D Pose Estimation for Any Object (CVPR 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.
