Research map / Label-efficient and robust learning / Domain adaptation and generalization
ADAS: A Direct Adaptation Strategy for Multi-Target Domain Adaptive Semantic Segmentation
Check 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 |
|---|---|---|---|
| Directly adapt one model to multiple targets using MTDT-Net and BARS rather than distilling pretrained single-target models | ADAS · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 47.9 -> 48.6 | Helped |
| Use ADAS rather than the strongest listed baseline | ADAS · GTA5 to Cityscapes, IDD, Mapillary (7 classes) | Avg. mIoU: 69.1 -> 71.3 | Helped |
| Use ADAS rather than CCL | ADAS · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 47.9 -> 48.6 | Helped |
| Use MTDT-Net instead of color transfer for domain image synthesis | ADAS segmentation training · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 37.8 -> 42.0 | Helped |
| Use MTDT-Net instead of DRANet for domain image synthesis | ADAS segmentation training · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 39.9 -> 42.0 | Helped |
| Add filtering of source-transfer labels with BARS (ablation) | ADAS without BARS filtering · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 42.0 -> 44.7 | Helped |
| Add filtering of target pseudo-labels with BARS (ablation) | ADAS without BARS filtering · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 42.0 -> 45.8 | Helped |
| Add both source-transfer label filtering and target pseudo-label filtering (ablation) | ADAS without BARS filtering · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 42.0 -> 48.6 | Helped |
| Use the target pseudo-label filter in addition to the source-transfer label filter (ablation) | ADAS with source-transfer filtering only · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 44.7 -> 48.6 | Helped |
| Use the source-transfer label filter in addition to the target pseudo-label filter (ablation) | ADAS with target pseudo-label filtering only · GTA5 to Cityscapes, IDD, Mapillary (19 classes) | Avg. mIoU: 45.8 -> 48.6 | Helped |
Closest papers by meaning
- Learning to Adapt via Latent Domains for Adaptive Semantic Segmentation (NeurIPS 2021 · similarity 0.91)
- Multi-Target Domain Adaptation With Collaborative Consistency Learning (CVPR 2021 · similarity 0.91)
- Multi-Source Domain Adaptation With Collaborative Learning for Semantic Segmentation (CVPR 2021 · similarity 0.91)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.90)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.89)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (CVPR 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.