Research map / Label-efficient and robust learning / Domain adaptation and generalization
DAST: Unsupervised Domain Adaptation in Semantic Segmentation Based on Discriminator Attention and Self-Training
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 |
|---|---|---|---|
| Use DAST instead of AdaptSegNet baseline | VGG-16 segmentation model · GTA5 → Cityscapes | mIoU: 35.0 -> 44.3 | Helped |
| Use DAST instead of AdaptSegNet baseline | ResNet-101 segmentation model · GTA5 → Cityscapes | mIoU: 41.4 -> 49.6 | Helped |
| Use DAST instead of FDA-MBT | ResNet-101 segmentation model · GTA5 → Cityscapes | mIoU: 50.5 -> 49.6 | Hurt |
| Use DAST instead of AdaptSegNet baseline | VGG-16 segmentation model · SYNTHIA → Cityscapes | mIoU*: 37.6 -> 46.5 | Helped |
| Use DAST instead of AdaptSegNet baseline | ResNet-101 segmentation model · SYNTHIA → Cityscapes | mIoU*: 45.9 -> 52.5 | Helped |
| Use DAST instead of FDA-MBT | ResNet-101 segmentation model · SYNTHIA → Cityscapes | mIoU*: 52.5 -> 52.5 | No clear effect |
| Add MSE adversarial loss (ablation) | AdaptSegNet baseline · GTA5 → Cityscapes | mIoU: 35.0 -> 35.9; 41.4 -> 43.4 | Helped |
| Use two discriminators for feature distribution alignment (ablation) | MSE-augmented AdaptSegNet · GTA5 → Cityscapes | mIoU: 35.9 -> 37.2; 43.4 -> 43.9 | Helped |
| Add attention to the two-discriminator setup (ablation) | MSE-augmented, two-discriminator AdaptSegNet · GTA5 → Cityscapes | mIoU: 37.2 -> 38.8; 43.9 -> 44.8 | Helped |
| Use self-training with q = 40% (ablation) | DAST · GTA5 → Cityscapes | mIoU: 49.0 | Helped |
| Use self-training with q = 50% rather than q = 40% (ablation) | DAST · GTA5 → Cityscapes | mIoU: 49.0 -> 49.6 | Helped |
| Use self-training with q = 60% rather than q = 50% (ablation) | DAST · GTA5 → Cityscapes | mIoU: 49.6 -> 48.7 | Hurt |
Closest papers by meaning
- Unsupervised Model Adaptation for Continual Semantic Segmentation (AAAI 2021 · similarity 0.91)
- Self-Ensembling Attention Networks: Addressing Domain Shift for Semantic Segmentation (AAAI 2019 · similarity 0.91)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.91)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.91)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.90)
- Towards Unsupervised Domain Bridging via Image Degradation in Semantic Segmentation (NeurIPS 2025 · similarity 0.90)
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.