Research map / Label-efficient and robust learning / Domain adaptation and generalization
Deliberated Domain Bridging for Domain Adaptive 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 |
|---|---|---|---|
| Replace direct adaptation with deliberated domain bridging using coarse and fine mixing paths plus cross-path distillation | ResNet-101, DeepLabv2 · GTA5→Cityscapes | mIoU: 60.8 -> 62.7 | Helped |
| Use deliberated domain bridging | ResNet-101, DeepLabv2 · GTA5+Synscapes→Cityscapes | mIoU: 59.0 -> 69.0 | Helped |
| Use deliberated domain bridging | ResNet-101, DeepLabv2 · GTA5→Cityscapes+Mapillary average | mIoU: 47.5 -> 58.6 | Helped |
| Combine coarse CutMix and fine ClassMix in mutually sampled bridging | Pseudo Labeling · GTA5→Cityscapes | mIoU: 51.7±0.4 -> 55.2±1.0 | Helped |
| Replace global interpolation Mixup with local replacement CutMix | Pseudo Labeling · GTA5→Cityscapes | mIoU: 31.6±0.6 -> 54.9±0.2 | Helped |
| Add the coarse region-level path (ablation) | DDB dual-path domain bridging · GTA5→Cityscapes | mIoU: 32.1 -> 56.5 | Helped |
| Add the fine class-level path (ablation) | DDB dual-path domain bridging · GTA5→Cityscapes | mIoU: 32.1 -> 58.2 | Helped |
| Use hard rather than soft distillation (ablation) | DDB cross-path knowledge distillation · GTA5→Cityscapes | mIoU: +0.9 mIoU | Helped |
| Use adaptive teacher ensembling (ablation) | DDB cross-path knowledge distillation · GTA5→Cityscapes | mIoU: +1.3 | Helped |
| Add hard distillation to the stage 1 student without adaptive ensembling (ablation) | DDB cross-path knowledge distillation · GTA5→Cityscapes | mIoU: 59.0 -> 59.9 | Helped |
| Add adaptive ensembling without hard distillation (ablation) | DDB cross-path knowledge distillation · GTA5→Cityscapes | mIoU: 59.0 -> 61.1 | Helped |
| Continue alternating optimization into a third round rather than stopping after round two (ablation) | DDB · GTA5→Cityscapes | mIoU: 62.7 -> 62.7 | No clear effect |
Closest papers by meaning
- Bidirectional Domain Mixup for Domain Adaptive Semantic Segmentation (AAAI 2023 · similarity 0.92)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.91)
- Towards Unsupervised Domain Bridging via Image Degradation in Semantic Segmentation (NeurIPS 2025 · similarity 0.91)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.90)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.90)
- Adapter Naturally Serves as Decoupler for Cross-Domain Few-Shot Semantic Segmentation (ICML 2025 · similarity 0.89)
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.