Research map / Label-efficient and robust learning / Domain adaptation and generalization
Rethinking Data Augmentation for Single-Source Domain Generalization in Medical Image Segmentation
Zixian Su, Kai Yao, Xi Yang, Kaizhu Huang, Qiufeng Wang, Jie Sun
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 SLAug instead of CSDG | U-Net with EfficientNet-b2 · Abdominal CT-MRI | Dice score (%): 86.31 -> 88.63 | Improved |
| Add SLAug instead of CSDG | U-Net with EfficientNet-b2 · Abdominal MRI-CT | Dice score (%): 80.40 -> 83.05 | Improved |
| Add SLAug instead of CSDG | U-Net with EfficientNet-b2 · Cardiac bSSFP-LGE | Dice score (%): 85.01 -> 86.69 | Improved |
| Add SLAug instead of RandConv | U-Net with EfficientNet-b2 · Cardiac LGE-bSSFP | Dice score (%): 87.24 -> 87.67 | Improved |
| Add GLA without LLA or SBF (ablation) | ERM · Abdominal CT-MRI | Dice score (%): 77.31 -> 80.28 | Improved |
| Add GLA without LLA or SBF (ablation) | ERM · Cardiac bSSFP-LGE | Dice score (%): 75.99 -> 74.21 | Hurt |
| Add LLA without GLA or SBF (ablation) | ERM · Abdominal CT-MRI | Dice score (%): 77.31 -> 75.43 | Hurt |
| Add LLA without GLA or SBF (ablation) | ERM · Cardiac bSSFP-LGE | Dice score (%): 75.99 -> 85.12 | Improved |
| Add both GLA and LLA without SBF (ablation) | ERM · Abdominal CT-MRI | Dice score (%): 77.31 -> 85.48 | Improved |
| Add both GLA and LLA without SBF (ablation) | ERM · Cardiac bSSFP-LGE | Dice score (%): 75.99 -> 85.55 | Improved |
| Add SBF between two GLA samples (ablation) | GLA-only augmentation · Abdominal CT-MRI | Dice score (%): 80.28 -> 83.57 | Improved |
| Add SBF between two LLA samples (ablation) | LLA-only augmentation · Abdominal CT-MRI | Dice score (%): 75.43 -> 75.42 | No clear effect |
Closest papers by meaning
- The Devil is in the Statistics: Mitigating and Exploiting Statistics Difference for Generalizable Semi-supervised Medical Image Segmentation (ECCV 2024 · similarity 0.90)
- Why Invariance is Not Enough for Biomedical Domain Generalization and How to Fix It (NeurIPS 2026 · similarity 0.89)
- LangDAug: Langevin Data Augmentation for Multi-Source Domain Generalization in Medical Image Segmentation (ICML 2025 · similarity 0.88)
- Learning Generalized Medical Image Segmentation from Decoupled Feature Queries (AAAI 2024 · similarity 0.88)
- Domain Generalized Medical Landmark Detection via Robust Boundary-Aware Pre-Training (AAAI 2025 · similarity 0.88)
- TinyMIG: Transferring Generalization from Vision Foundation Models to Single-Domain Medical Imaging (ICML 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.
