Research map / Label-efficient and robust learning / Recognition and long-tail
A2B: Anchor to Barycentric Coordinate for Robust Correspondence
Weiyue Zhao, Hao Lü, Zhiguo Cao, Xin Li
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 barycentric coordinates from seed correspondences | SuperGlue · PhotoTourism (HardNet) | AUC@20: 53.45 -> 58.94 | Improved |
| add barycentric coordinates from seed correspondences | SuperGlue · PhotoTourism (RootSIFT) | AUC@20: 76.58 -> 77.98 | Improved |
| add barycentric coordinates to SuperGlue with HardNet | SuperGlue · Outdoor PhotoTourism (HardNet) | precision, recall, pose: +6.3% in precision, +11.0% in recall and +5.5% in pose estimation | Improved |
| add barycentric coordinates to SuperGlue | SuperGlue · SUN3D (RootSIFT) | AUC@5: 13.04 -> 12.89 | No clear effect |
| add barycentric coordinates to OANet with RootSIFT | OANet · YFCC100M (RootSIFT) | precision, recall: 46.10 -> 59.30 | Improved |
| add barycentric coordinates to OANet with SuperPoint | OANet · YFCC100M (SuperPoint) | AUC@20: 67.93 -> 67.63 | No clear effect |
| add DEGREE to OANet | OANet · Oxford | F-measure: 95.86 -> 97.67 | Improved |
| add NM-block to SMFNet (ablation) | SMFNet · PhotoTourism seed correspondences | Top-10: 8.91 -> 9.88 | Improved |
| add AR-block to SMFNet (ablation) | SMFNet · PhotoTourism seed correspondences | Top-10: 8.73 -> 9.88 | Improved |
| add fine matching to SMFNet (ablation) | SMFNet · PhotoTourism seed correspondences | Top-10: 6.17 -> 9.88 | Improved |
| add position encoding in coarse matching (ablation) | SMFNet · PhotoTourism seed correspondences | Top-10: 9.88 -> 9.91 | No clear effect |
| increase seed correspondences K from 5 to 6 (ablation) | DEGREE+SuperGlue · YFCC100M (HardNet) | pose accuracy | No clear effect |
Closest papers by meaning
- Learning Probabilistic Coordinate Fields for Robust Correspondences (TPAMI 2023 · similarity 0.89)
- OANet: Learning Two-View Correspondences and Geometry Using Order-Aware Network (TPAMI 2020 · similarity 0.85)
- Learning Two-View Correspondences and Geometry Using Order-Aware Network (ICCV 2019 · similarity 0.85)
- Scalable and Generalizable Correspondence Pruning via Geometry-Consistent Pre-Training (TPAMI 2026 · similarity 0.85)
- Learning Affine Correspondences by Integrating Geometric Constraints (CVPR 2025 · similarity 0.85)
- DGC-GNN: Leveraging Geometry and Color Cues for Visual Descriptor-Free 2D-3D Matching (CVPR 2024 · similarity 0.83)
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.
