Research map / Graph learning / Graph algorithms and clustering
Deep Learning of Graph Matching
Read the paperCheck 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 |
|---|---|---|---|
| refine the graph-matching network end-to-end instead of using its initial feature weights (ablation) | GMNwVGG on MPI-Sintel · MPI-Sintel test | PCK@10 pixels: 88.03 -> 92.6 | Helped |
| use graph matching with initial VGG features rather than nearest-neighbor matching | VGG-16 features on MPI-Sintel · MPI-Sintel test | PCK@10 pixels: 85.9 -> 88.03 | Helped |
| refine the graph-matching network end-to-end instead of using its initial feature weights (ablation) | GMNwVGG on CUB-200-2011 · CUB-200-2011 test | EPE (in pixels): 41.63 -> 17.02 | Helped |
| refine the graph-matching network end-to-end instead of using its initial feature weights (ablation) | GMNwVGG on CUB-200-2011 · CUB-200-2011 test | PCK@0.05: 0.63 -> 0.86 | Helped |
| add the Voting layer at test time (ablation) | GMNwVGG-T on CUB-200-2011 · CUB-200-2011 test | EPE (in pixels): 18.22 -> 17.02 | Helped |
| add the Voting layer at test time (ablation) | GMNwVGG-T on CUB-200-2011 · CUB-200-2011 test | PCK@0.05: 0.83 -> 0.86 | Helped |
| use graph matching with initial VGG features rather than nearest-neighbor matching | VGG-16 features on CUB-200-2011 · CUB-200-2011 test | EPE (in pixels): 59.05 -> 41.63 | Helped |
| refine the graph-matching network end-to-end instead of using its initial feature weights (ablation) | GMNwVGG on PASCAL VOC keypoints · PASCAL VOC keypoints test | PCK@0.1 (Class average): 29.8 -> 40.6 | Helped |
| use graph matching with initial VGG features rather than nearest-neighbor matching | VGG-16 features on PASCAL VOC keypoints · PASCAL VOC keypoints test | PCK@0.1 (Class average): 25.4 -> 29.8 | Helped |
| use end-to-end graph matching rather than UCN | keypoint matching on PASCAL VOC · PASCAL VOC keypoints test | PCK@0.1 (Class average): 38.9 -> 40.6 | Helped |
| use end-to-end graph matching rather than UCN | matching on MPI-Sintel · MPI-Sintel test | PCK@10 pixels: 91.5 -> 92.6 | Helped |
Closest papers by meaning
- Learning Combinatorial Solver for Graph Matching (CVPR 2020 · similarity 0.92)
- Deep Graph Matching Under Quadratic Constraint (CVPR 2021 · similarity 0.91)
- Learning deep graph matching with channel-independent embedding and Hungarian attention (ICLR 2020 · similarity 0.91)
- Deep Latent Graph Matching (ICML 2021 · similarity 0.91)
- Learning Partial Graph Matching via Optimal Partial Transport (ICLR 2025 · similarity 0.90)
- Discrete Cycle-Consistency Based Unsupervised Deep Graph Matching (AAAI 2024 · 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.