Research map / Label-efficient and robust learning / Domain adaptation and generalization
Undoing the Damage of Label Shift for Cross-Domain Semantic Segmentation
Yahao Liu, Jinhong Deng, Jiale Tao, Tong Chu, Lixin Duan, Wen Li
Check 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 |
|---|---|---|---|
| Apply inference adjustment after conditional alignment | FADA · GTA5 to Cityscapes | mIoU: 46.9 -> 48.8 | Improved |
| Apply classifier refinement after conditional alignment | FADA · GTA5 to Cityscapes | mIoU: 46.9 -> 49.0 | Improved |
| Add classifier rectification | FADA · SYNTHIA to Cityscapes | mIoU: 40.9 -> 42.5; 40.9 -> 43.0 | Improved |
| Add classifier rectification to ProDA | ProDA · GTA5 to Cityscapes | mIoU: 57.5 -> 59.3; 1.8 | Improved |
| Add classifier rectification to ProDA | ProDA · SYNTHIA to Cityscapes | mIoU: 55.5 -> 56.7 | Improved |
| Add classifier rectification to ProDA | ProDA · SYNTHIA to Cityscapes | mIoU* (13 classes): 62.0 -> 64.5 | Improved |
| Add classifier rectification to IAST | IAST · GTA5 to Cityscapes | mIoU: 51.5 -> 55.5; 53.0 -> 55.5; 2.5% | Improved |
| Add classifier rectification to self-distillation | FADA+SD · GTA5 to Cityscapes | mIoU: 49.2 -> 52.7; 3.5 | Improved |
| Use inference adjustment with Modified ASPP instead of Original ASPP (ablation) | Our IA · GTA5 to Cityscapes | mIoU: 48.8 -> 49.2 | Improved |
| Use classifier refinement with Modified ASPP instead of Original ASPP (ablation) | Our CR · GTA5 to Cityscapes | mIoU: 49.0 -> 48.7 | Hurt |
Closest papers by meaning
- TACS: Taxonomy Adaptive Cross-Domain Semantic Segmentation (ECCV 2022 · similarity 0.89)
- Towards Unbiased Learning in Semi-Supervised Semantic Segmentation (ICLR 2025 · similarity 0.88)
- DynAlign: Unsupervised Dynamic Taxonomy Alignment for Cross-Domain Segmentation (ICLR 2025 · similarity 0.88)
- Connectivity-Driven Pseudo-Labeling Makes Stronger Cross-Domain Segmenters (NeurIPS 2024 · similarity 0.87)
- Generalize or Detect? Towards Robust Semantic Segmentation Under Multiple Distribution Shifts (NeurIPS 2024 · similarity 0.87)
- Heterogeneous Label Shift: Theory and Algorithm (ICML 2025 · similarity 0.87)
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.
