Research map / Label-efficient and robust learning / Domain adaptation and generalization
Joint Adversarial Learning for Domain Adaptation in Semantic Segmentation
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 JAL (low-level discriminator loss on high-level decoder) to two-level adaptation (ablation) | AdaptSegNet · GTA5->Cityscapes VGG16 | mIoU: 35.0 -> 36.7 | Helped |
| add JAL (without WTM) (ablation) | AdaptSegNet · GTA5->Cityscapes ResNet101 | mIoU: 42.4 -> 41.5 | Hurt |
| add WTM on both decoders (no JAL) (ablation) | AdaptSegNet · GTA5->Cityscapes VGG16 | mIoU: 35.0 -> 36.4 | Helped |
| add WTM on both decoders (no JAL) (ablation) | AdaptSegNet · GTA5->Cityscapes ResNet101 | mIoU: 42.4 -> 42.9 | Helped |
| add WTM on top of JAL (ablation) | JAL · GTA5->Cityscapes ResNet101 | mIoU: 41.5 -> 43.5 | Helped |
| add WTM on top of JAL (ablation) | JAL · GTA5->Cityscapes VGG16 | mIoU: 36.7 -> 38.5 | Helped |
| WTM on Dec1 only vs none (ablation) | AdaptSegNet · GTA5->Cityscapes VGG16 | mIoU: 35.0 -> 35.9 | Helped |
| widen discriminator (width 2 or 4) instead of JAL (ablation) | AdaptSegNet · GTA5->Cityscapes VGG16 | mIoU: 35.0 -> 35.6 | Helped |
| use multiple high-level discriminators (N=2 to 4) (ablation) | AdaptSegNet · GTA5->Cityscapes VGG16 | mIoU: 35.0 -> 35.5 | Helped |
| add ASPP module to discriminator (ablation) | AdaptSegNet · GTA5->Cityscapes VGG16 | mIoU: 35.0 -> 35.3 | No clear effect |
| add JAL+WTM to Bidirectional Learning with style translation | Bidirectional Learning · GTA5->Cityscapes ST iter 1 | mIoU: 42.7 -> 43.8 | Helped |
| add JAL+WTM to Bidirectional Learning with ST+SSL | Bidirectional Learning · GTA5->Cityscapes ST+SSL iter 2 | mIoU: 48.5 -> 49.3 | Helped |
Closest papers by meaning
- An Adversarial Perturbation Oriented Domain Adaptation Approach for Semantic Segmentation (AAAI 2020 · similarity 0.92)
- Learning to Adapt Structured Output Space for Semantic Segmentation (CVPR 2018 · similarity 0.91)
- DAST: Unsupervised Domain Adaptation in Semantic Segmentation Based on Discriminator Attention and Self-Training (AAAI 2021 · similarity 0.91)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.89)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.88)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 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.