Research map / Detection and segmentation / Weakly supervised detection and segmentation
Recurrent Transformer Networks for Semantic Correspondence
Seungryong Kim, Stephen Lin, SANG RYUL JEON, Dongbo Min, Kwanghoon Sohn
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 |
|---|---|---|---|
| Use recurrent geometric matching with ResNet features instead of GMat. w/Inl. | Semantic correspondence · TSS average | matching accuracy: 0.737 -> 0.772 | Improved |
| Use RTNs with ResNet instead of the strongest listed baseline | Semantic correspondence · PF-WILLOW | PCK α=0.05: 0.386 -> 0.413 | Improved |
| Use RTNs with ResNet instead of the strongest listed baseline | Semantic correspondence · PF-WILLOW | PCK α=0.1: 0.704 -> 0.719 | Improved |
| Use RTNs with ResNet instead of the strongest listed baseline | Semantic correspondence · PF-PASCAL | PCK α=0.05: 0.490 -> 0.552 | Improved |
| Use RTNs with ResNet instead of the strongest listed baseline | Semantic correspondence · PF-PASCAL | PCK α=0.1: 0.748 -> 0.759 | Improved |
| Use RTNs with ResNet instead of the strongest listed baseline | Semantic correspondence · PF-PASCAL | PCK α=0.15: 0.840 -> 0.852 | Improved |
| Use 3–5 recurrent iterations (ablation) | RTNs w/ResNet · TSS | matching accuracy | Improved |
| Enlarge the local search window up to 9×9 (ablation) | RTNs · TSS | matching accuracy | Improved |
| Enlarge the local search window beyond 9×9 (ablation) | RTNs · TSS | matching accuracy | Hurt |
| Use ResNet features instead of VGGNet or CAT-FCSS features (ablation) | RTNs · TSS | matching accuracy | Improved |
| Jointly learn the feature extractor with the geometric matching network instead of using a pretrained backbone (ablation) | RTNs · TSS | matching accuracy | Improved |
Closest papers by meaning
- Learning Semantic Correspondence Exploiting an Object-Level Prior (TPAMI 2020 · similarity 0.89)
- Pyramidal Semantic Correspondence Networks (TPAMI 2021 · similarity 0.89)
- SFNet: Learning Object-Aware Semantic Correspondence (CVPR 2019 · similarity 0.88)
- Universal Correspondence Network (NeurIPS 2016 · similarity 0.88)
- PARN: Pyramidal Affine Regression Networks for Dense Semantic Correspondence (ECCV 2018 · similarity 0.88)
- Semantic Correspondence: Unified Benchmarking and a Strong Baseline (TPAMI 2025 · 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.
