Research map / Graph learning / Graph algorithms and clustering
Learning Combinatorial Embedding Networks for Deep Graph Matching
Runzhong Wang, Junchi Yan, Xiaokang Yang
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 |
|---|---|---|---|
| replace offset loss with permutation loss (ablation) | GMN · Pascal VOC Keypoint | matching accuracy (%): 55.3 -> 57.9 | Improved |
| use permutation loss instead of offset loss (ablation) | PIA-GM · Pascal VOC Keypoint | matching accuracy (%): 61.6 -> 63.0 | Improved |
| add cross-graph embedding on top of intra-graph (PCA-GM vs PIA-GM) | PIA-GM · Pascal VOC Keypoint | matching accuracy (%): 63.0 -> 63.8 | Improved |
| add cross-graph embedding (ablation) | PCA-GM · Pascal VOC Keypoint (ablation) | matching accuracy (%): 62.1 -> 63.6 | Improved |
| add learnable affinity metric (ablation) | PCA-GM · Pascal VOC Keypoint (ablation) | matching accuracy (%): 63.6 -> 63.8 | Improved |
| add intra-graph embedding (ablation) | PCA-GM · Pascal VOC Keypoint (ablation) | matching accuracy (%): 54.8 -> 62.1 | Improved |
| fine-tune VGG16 features on graph matching data (ablation) | PCA-GM · Pascal VOC Keypoint (ablation) | matching accuracy (%): 41.9 -> 54.8 | Improved |
| predict S iteratively in cross-graph module (7 iterations) instead of from shallow layer (ablation) | PCA-GM · Pascal VOC Keypoint | matching accuracy (%): 63.8 -> 46.2 | Hurt |
| predict S iteratively with 1 iteration instead of simple design (ablation) | PCA-GM · Pascal VOC Keypoint | matching accuracy (%): 63.8 -> 63.1 | Hurt |
| fine-tune on Willow after VOC pretraining | PCA-GM · Willow ObjectClass (duck) | matching accuracy (%): 82.4 -> 93.5 | Improved |
| use PCA-GM instead of GMN | PCA-GM-VOC · Willow ObjectClass (wine bottle) | matching accuracy (%): 70.5 -> 95.1 | Improved |
| use PCA-GM instead of GMN (motorbike, Willow-trained) | PCA-GM-Willow · Willow ObjectClass (motorbike) | matching accuracy (%): 71.4 -> 76.7 | Improved |
Closest papers by meaning
- Combinatorial Learning of Robust Deep Graph Matching: An Embedding Based Approach (TPAMI 2020 · similarity 0.97)
- Learning deep graph matching with channel-independent embedding and Hungarian attention (ICLR 2020 · similarity 0.94)
- Deep Learning of Graph Matching (CVPR 2018 · similarity 0.93)
- Learning Combinatorial Solver for Graph Matching (CVPR 2020 · similarity 0.93)
- Learning Partial Graph Matching via Optimal Partial Transport (ICLR 2025 · similarity 0.91)
- 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.
