Research map / 3D vision / Human pose estimation
GS-Pose: Category-Level Object Pose Estimation via Geometric and Semantic Correspondence
Pengyuan Wang, Takuya Ikeda, Robert Lee, Koichi Nishiwaki
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 transformer feature fusion network instead of directly matching raw 3D semantic features (ablation) | GS-Pose · NOCS REAL275 | 3D25 / 3D50 (%): 2.2 -> full model; 0.1 -> full model | Improved |
| add inlier probability prediction to matching (ablation) | GS-Pose · NOCS REAL275 | 3D25 / 3D50: drop by 17.2% and 14.3% without it | Improved |
| add inlier probability prediction to matching (ablation) | GS-Pose · NOCS REAL275 | 5/10/15 deg 5cm: drop by 8.9%, 10.8%, 6.7% without it | Improved |
| add symmetry handling (disambiguating ground truth) in training (ablation) | GS-Pose · NOCS REAL275 | 3D50: decreases greatly to 44.5% without it | Improved |
| add symmetry handling in training (ablation) | GS-Pose · NOCS REAL275 | 5/10/15 deg 5cm: drop by 16.1%, 19.6%, 16.9% without it | Improved |
| increase number of synthetic training objects for CPPF (ablation) | CPPF · NOCS REAL275 | 5/10/15 deg 5cm | Improved |
| increase synthetic training objects from 10 to 20 and 40 (ablation) | GS-Pose · NOCS REAL275 | 3D25 / 3D50 | No clear effect |
| use GS-Pose (synthetic only) instead of CPPF (synthetic only) | NOCS REAL275 · NOCS REAL275 | 3D50: +36.8% | Improved |
| use GS-Pose instead of CPPF | SUN RGB-D chairs · SUN RGB-D | 3D10 / 3D25: +20.8% and +18.4% | Improved |
| GS-Pose trained on synthetic data versus DualPoseNet trained on real data | Wild6D · Wild6D | 10 deg 5cm: +8.5% | Improved |
Closest papers by meaning
- GCE-Pose: Global Context Enhancement for Category-level Object Pose Estimation (CVPR 2025 · similarity 0.90)
- SchemaPose: RGB-Based Category-Level Pose Estimation with Parametric Category Schema (NeurIPS 2026 · similarity 0.88)
- SecondPose: SE(3)-Consistent Dual-Stream Feature Fusion for Category-Level Pose Estimation (CVPR 2024 · similarity 0.88)
- Vision Foundation Model Enables Generalizable Object Pose Estimation (NeurIPS 2024 · similarity 0.88)
- Universal Features Guided Zero-Shot Category-Level Object Pose Estimation (AAAI 2025 · similarity 0.88)
- GenCOPE: Syn2Real Generalized Category-Level Object Pose Estimation for Robotic Picking (NeurIPS 2026 · 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.
