Research map / Label-efficient and robust learning / Domain adaptation and generalization
MIC: Masked Image Consistency for Context-Enhanced Domain Adaptation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 Masked Image Consistency to HRDA | HRDA · GTA→Cityscapes | mIoU: 73.8 -> 75.9 | Helped |
| Add Masked Image Consistency to HRDA | HRDA · Cityscapes→DarkZurich | mIoU: +4.3 mIoU | Helped |
| Add Masked Image Consistency to SDAT | SDAT with ViT-B/16 · VisDA-2017 | mean accuracy: +3.0 percent points | Helped |
| Add Masked Image Consistency to SADA | SADA · Cityscapes→Foggy Cityscapes | mAP: +3.6 mAP | Helped |
| Add image patch masking (ablation) | MIC with DAFormer · GTA→Cityscapes | mIoU: 50.6 -> 70.6 | Helped |
| Add color augmentation (ablation) | MIC with DAFormer · GTA→Cityscapes | mIoU: 70.3 -> 70.6 | No clear effect |
| Use an EMA teacher instead of regular model predictions (ablation) | MIC with DAFormer · GTA→Cityscapes | mIoU: 69.9 -> 70.6 | Helped |
| Add pseudo-label confidence weighting (ablation) | MIC with DAFormer · GTA→Cityscapes | mIoU: 69.0 -> 70.6 | Helped |
| Apply MIC to source images rather than target images (ablation) | HRDA · GTA→Cityscapes | mIoU: 73.8 -> 71.1 | Hurt |
| Apply MIC to source images as well as target images (ablation) | HRDA · Cityscapes→ACDC | mIoU: 65.3 -> 68.0 | Helped |
| Calculate MIC loss on both masked and unmasked patches (ablation) | MIC with DAFormer · GTA→Cityscapes | mIoU: 69.8 -> 70.6 | Helped |
| Use masked image modeling as additional pretraining (ablation) | SDAT · VisDA-2017 | mean accuracy: -1.4 | Hurt |
Closest papers by meaning
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.87)
- Mean Teacher DETR with Masked Feature Alignment: A Robust Domain Adaptive Detection Transformer Framework (AAAI 2024 · similarity 0.87)
- Enhancing Domain Adaptation through Prompt Gradient Alignment (NeurIPS 2024 · similarity 0.86)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.86)
- Toward Improving Robustness and Accuracy in Unsupervised Domain Adaptation (AAAI 2025 · similarity 0.86)
- Universal Domain Adaptive Object Detection via Dual Probabilistic Alignment (AAAI 2025 · 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.