Research map / Detection and segmentation / CNN features
Dynamic Context Correspondence Network for Semantic Alignment
Shuaiyi Huang, Qiuyue Wang, Songyang Zhang, Shipeng Yan, Xuming He
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 spatial context encoder (kernel 25) (ablation) | NC-Net · PF-PASCAL | PCK alpha=0.1 (%): 78.9 -> 80.9 | Improved |
| add dynamic fusion with correlation embedding (no aux loss) on top of context encoder (ablation) | Baseline+S · PF-PASCAL | PCK alpha=0.1 (%): 80.9 -> 81.0 | No clear effect |
| add multi-auxiliary task loss (ablation) | Baseline+SD · PF-PASCAL | PCK alpha=0.1 (%): 81.0 -> 82.3 | Improved |
| use dynamic attention fusion instead of average fusion (ablation) | Baseline+SAA · PF-PASCAL | PCK alpha=0.1 (%): 80.2 -> 82.3 | Improved |
| add correlation map embedding to dynamic fusion (ablation) | Baseline+SCA · PF-PASCAL | PCK alpha=0.1 (%): 79.9 -> 82.3 | Improved |
| average fusion of correlation maps instead of using context feature alone (ablation) | Baseline+S · PF-PASCAL | PCK alpha=0.1 (%): 80.9 -> 80.2 | Hurt |
| increase context kernel size from 11 to 25 (ablation) | Baseline+S · PF-PASCAL | PCK alpha=0.1 (%): 78.9 -> 80.9 | Improved |
| increase context kernel size from 25 to 31 (ablation) | Baseline+S · PF-PASCAL | PCK alpha=0.1 (%): 80.9 -> 77.1 | Hurt |
| add dynamic fusion and aux loss at kernel size 31 (ablation) | Baseline+S · PF-PASCAL | PCK alpha=0.1 (%): 77.1 -> 80.7 | Improved |
| overall DCCNet vs NC-Net | NC-Net · PF-PASCAL | PCK alpha=0.1 (%): 78.9 -> 82.3 | Improved |
| DCCNet vs NC-Net at stricter threshold | NC-Net · PF-WILLOW | PCK alpha=0.05 (%): 44.0 -> 43.6 | Hurt |
| DCCNet vs NC-Net on PF-WILLOW at alpha=0.10 | NC-Net · PF-WILLOW | PCK alpha=0.10 (%): 72.7 -> 73.8 | Improved |
Closest papers by meaning
- Semantic Correspondence: Unified Benchmarking and a Strong Baseline (TPAMI 2025 · similarity 0.90)
- Learning Semantic Correspondence Exploiting an Object-Level Prior (TPAMI 2020 · similarity 0.89)
- End-to-End Weakly-Supervised Semantic Alignment (CVPR 2018 · similarity 0.88)
- Multi-Scale Matching Networks for Semantic Correspondence (ICCV 2021 · similarity 0.88)
- SFNet: Learning Object-Aware Semantic Correspondence (CVPR 2019 · similarity 0.88)
- Pixel-level Semantic Correspondence through Layout-aware Representation Learning and Multi-scale Matching Integration (CVPR 2024 · similarity 0.86)
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.
