Research map / 3D vision / Human pose estimation
GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object 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 |
|---|---|---|---|
| Use geometry-guided direct pose regression rather than correspondence-only pose estimation | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 89.9 -> 93.7 | Helped |
| Use GDR-Net with one pose estimator per object instead of one shared estimator | GDR-Net · YCB-V | ADD(-S) (%): 49.1 -> 60.1 | Helped |
| Add geometric supervision LGeom (ablation) | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 92.2 -> 93.7 | Helped |
| Add dense correspondence guidance M2D (ablation) | GDR-Net without geometric loss on LM · LM | ADD(-S) 0.1d (%): 85.6 -> 92.2 | Helped |
| Use Patch-PnP instead of PointNet-like PnP (ablation) | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 92.3 -> 93.7 | Helped |
| Use Patch-PnP instead of BPnP (ablation) | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 92.0 -> 93.7 | Helped |
| Use allocentric R6d instead of egocentric R6d (ablation) | GDR-Net on LM · LM | 2°, 2 cm (%): 60.4 -> 62.1 | Helped |
| Use allocentric R6d instead of allocentric quaternion (ablation) | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 90.5 -> 93.7 | Helped |
| Use allocentric R6d instead of allocentric log quaternion (ablation) | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 88.9 -> 93.7 | Helped |
| Use allocentric R6d instead of allocentric Lie algebra vector (ablation) | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 89.7 -> 93.7 | Helped |
| Use scale-invariant tSITE instead of directly regressing translation t (ablation) | GDR-Net on LM · LM | ADD(-S) 0.1d (%): 92.4 -> 93.7 | Helped |
| Use scale-invariant tSITE instead of object center and absolute distance (ablation) | GDR-Net on LM · LM | 2°, 2 cm (%): 50.4 -> 62.1 | Helped |
Closest papers by meaning
- Single-Stage 6D Object Pose Estimation (CVPR 2020 · similarity 0.90)
- ES6D: A Computation Efficient and Symmetry-Aware 6D Pose Regression Framework (CVPR 2022 · similarity 0.90)
- One2Any: One-Reference 6D Pose Estimation for Any Object (CVPR 2025 · similarity 0.87)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.87)
- FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · similarity 0.87)
- GIVEPose: Gradual Intra-class Variation Elimination for RGB-based Category-Level Object Pose Estimation (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.