Research map / Label-efficient and robust learning / Domain adaptation and generalization
Unified Domain Adaptive Semantic Segmentation
Zhe Zhang, Gaochang Wu, Jing Zhang, Xiatian Zhu, Dacheng Tao, Tianyou Chai
Read the paperCheck 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 |
|---|---|---|---|
| add video patch template mixing (ablation) | TPS-style baseline · Synthia-Seq to Cityscapes-Seq | mIoU: 52.01 -> 55.71 | Improved |
| add feature-level template mixing on top of V-template (ablation) | QuadMix · Synthia-Seq to Cityscapes-Seq | mIoU: 55.71 -> 57.25 | Improved |
| add optical-guided video feature aggregation (ablation) | TPS-style baseline · Synthia-Seq to Cityscapes-Seq | mIoU: 52.01 -> 54.01 | Improved |
| combine V-template, F-template and aggregation (ablation) | TPS-style baseline · Synthia-Seq to Cityscapes-Seq | mIoU: 52.01 -> 59.62 | Improved |
| add warm-up initialization to the full model (ablation) | QuadMix · Synthia-Seq to Cityscapes-Seq | mIoU: 59.62 -> 61.48 | Improved |
| add F-template when V-template, Agg and warm-up are present (ablation) | QuadMix · Synthia-Seq to Cityscapes-Seq | mIoU: 60.26 -> 61.48 | Improved |
| add warm-up to aggregation-only model (ablation) | Agg only · Synthia-Seq to Cityscapes-Seq | mIoU: 54.01 -> 55.97 | Improved |
| add warm-up to V+F model (ablation) | QuadMix without Agg · Synthia-Seq to Cityscapes-Seq | mIoU: 57.25 -> 57.95 | Improved |
| use video patch template instead of image patch template (ablation) | V-template variant · Synthia-Seq to Cityscapes-Seq | mIoU: drop of 1.37 | Improved |
| switch backbone from DeepLab-V2 to transformer (ablation) | QuadMix · Synthia-Seq to Cityscapes-Seq | mIoU: 59.6 -> 67.2 | Improved |
| switch backbone from DeepLab-V2 to transformer | QuadMix · VIPER to Cityscapes-Seq | mIoU: 54.9 -> 58.2 | Improved |
| switch backbone from DeepLab-V2 to transformer | QuadMix · GTAV to Cityscapes | mIoU: 66.8 -> 76.1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DAFormer — DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation (CVPR 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.95)
- I2F: A Unified Image-to-Feature Approach for Domain Adaptive Semantic Segmentation (TPAMI 2022 · similarity 0.93)
- Towards Unsupervised Domain Bridging via Image Degradation in Semantic Segmentation (NeurIPS 2025 · similarity 0.92)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.91)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.91)
- DynAlign: Unsupervised Dynamic Taxonomy Alignment for Cross-Domain Segmentation (ICLR 2025 · similarity 0.91)
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.
