Research map / Label-efficient and robust learning / Domain adaptation and generalization
FDA: Fourier Domain Adaptation for 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 |
|---|---|---|---|
| Apply Fourier low-frequency amplitude transfer and multi-band self-training | DeepLabV2 ResNet101 · GTA5→CityScapes | mIoU: 50.45 vs 48.5 | Helped |
| Apply Fourier low-frequency amplitude transfer and multi-band self-training | VGG16 · GTA5→CityScapes | mIoU: 42.2 vs 41.3 | Helped |
| Apply Fourier low-frequency amplitude transfer and multi-band self-training | ResNet101 · SYNTHIA→CityScapes (13 classes) | mIoU: 52.5 vs 51.4 | Helped |
| Apply Fourier low-frequency amplitude transfer and multi-band self-training | VGG16 · SYNTHIA→CityScapes (16 classes) | mIoU: 40.5 vs 39.0 | Helped |
| Vary the FDA spectral neighborhood size among β=0.01, 0.05, and 0.09 (ablation) | Single-scale FDA · GTA5→CityScapes | mIoU: 44.61; 44.6; 45.01 | No clear effect |
| Add entropy regularization to single-scale FDA (ablation) | FDA with β=0.09 · GTA5→CityScapes | mIoU: 44.64 -> 45.01 | Helped |
| Add single-network self-supervised training without averaging (ablation) | FDA with β=0.09 · GTA5→CityScapes | mIoU: 45.01 -> 45.42 | Helped |
| Average predictions from different β models (ablation) | FDA models trained from scratch · GTA5→CityScapes | mIoU: 45.01 -> 46.77 | Helped |
| Add the first round of self-supervised training to multi-band prediction averaging (ablation) | MBT · GTA5→CityScapes | mIoU: 46.77 -> 48.14 | Helped |
| Add the second round of self-supervised training to multi-band prediction averaging (ablation) | MBT · GTA5→CityScapes | mIoU: 48.14 -> 50.45 | Helped |
| Use single-network self-training rather than averaging multiple β-model predictions (ablation) | FDA with β=0.09 · GTA5→CityScapes | mIoU: 45.01 -> 45.42; 45.42 vs 46.77 | Mixed |
Closest papers by meaning
- Joint Adversarial Learning for Domain Adaptation in Semantic Segmentation (AAAI 2020 · similarity 0.91)
- Source-Free Domain Adaptation for Semantic Segmentation (CVPR 2021 · similarity 0.91)
- Learning to Adapt via Latent Domains for Adaptive Semantic Segmentation (NeurIPS 2021 · similarity 0.91)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.90)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.90)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · 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.