Research map / Label-efficient and robust learning / Domain adaptation and generalization
Geometric Correspondence Constrained Pseudo-Label Alignment for Source-Free Domain Adaptive Fundus Image Segmentation
Zhouhongyuan Hu, Lei Zhang, Lituan Wang, Zhenwei Zhang, Minjuan Zhu, Zhenbin Wang
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 pseudo-label alignment alone without SAPE (ablation) | GCC · Drishti-GS to RIM-ONE-r3 | Avg Dice (%): 84.69 -> 84.01 | Hurt |
| add Gaussian perturbation to pseudo-label refinement (ablation) | Refined pseudo-label · Drishti-GS to RIM-ONE-r3 train set | Optic cup Dice (%): 69.29 -> 72.34 | Improved |
| add SAPE and filter to alignment (full GCC) (ablation) | GCC · Drishti-GS to RIM-ONE-r3 | Avg Dice (%): 84.01 -> 85.87 | Improved |
| add semantic loss Lsem (ablation) | geometric correspondence learning · Drishti-GS to RIM-ONE-r3 | Avg Dice (%): 84.89 -> 85.15 | Improved |
| add semantic loss Lsem (ablation) | geometric correspondence learning · Drishti-GS to RIM-ONE-r3 | Avg ASD (pixel): 8.13 -> 7.36 | Improved |
| add entropy loss Lent alone (ablation) | geometric correspondence learning · Drishti-GS to RIM-ONE-r3 | Avg Dice (%): 84.89 -> 84.80 | No clear effect |
| add entropy loss Lent alone (ablation) | geometric correspondence learning · Drishti-GS to RIM-ONE-r3 | Avg ASD (pixel): 8.13 -> 7.27 | Improved |
| combine Lsem and Lent (ablation) | geometric correspondence learning · Drishti-GS to RIM-ONE-r3 | Avg Dice (%): 84.89 -> 85.87 | Improved |
| combine Lsem and Lent (ablation) | geometric correspondence learning · Drishti-GS to RIM-ONE-r3 | Avg ASD (pixel): 8.13 -> 6.71 | Improved |
| use sample-adaptive perturbation (SAPE) instead of fixed perturbation ratios (ablation) | GCC · Drishti-GS to RIM-ONE-r3 | Avg Dice (%) | Improved |
| vary initial low-Q proportion alpha (ablation) | GCC · Drishti-GS to RIM-ONE-r3 | Avg Dice (%) | No clear effect |
| use GCC instead of CPR | GCC · Drishti-GS to RIM-ONE-r3 | Avg Dice (%): 84.61 -> 85.87 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TENT — Tent: Fully Test-Time Adaptation by Entropy Minimization (ICLR 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Connectivity-Driven Pseudo-Labeling Makes Stronger Cross-Domain Segmenters (NeurIPS 2024 · similarity 0.86)
- Why Invariance is Not Enough for Biomedical Domain Generalization and How to Fix It (NeurIPS 2026 · similarity 0.85)
- Leveraging Anatomical Consistency for Multi-Object Detection in Ultrasound Images via Source-free Unsupervised Domain Adaptation (AAAI 2025 · similarity 0.85)
- MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling (CVPR 2024 · similarity 0.85)
- A Unified Degradation-Robust Approach to SSL and UDA for 3D Medical Images (AAAI 2025 · similarity 0.84)
- Constructing and Exploring Intermediate Domains in Mixed Domain Semi-supervised Medical Image Segmentation (CVPR 2024 · similarity 0.84)
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.
