Object Detection on Cityscapes-C (test)
45.7mAP (Clean)Oracle - Train on target
Evaluation results
| Method | Links | |||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Oracle - Train on targetTraining Source=C2026.01 | 45.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 43.7 | 59.9 | 58.6 | 39.1 | 39.8 | 53.1 | 34.5 | 36.9 | |
| GB-DAL+Div+SNFTraining Source=F&B, Augmentation-free=false2026.01 | 45.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 44.1 | 55.3 | 58.2 | 38.8 | 41.2 | 51 | 31.5 | 41.5 | |
| GB-DAL+SNFTraining Source=F&B, Augmentation-free=false2026.01 | 44.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 43.7 | 55.1 | 57.8 | 38.5 | 41.3 | 51.1 | 30.8 | 40.8 | |
| DivAlign+SNFTraining Source=F&B, Augmentation-free=false2026.01 | 44.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 41.8 | 54.9 | 58.8 | 37.2 | 41.3 | 48.3 | 32.3 | 42.8 | |
| DivAlignTraining Source=F&B, Augmentation-free=false2026.01 | 44.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 41.8 | 54.7 | 59.1 | 35.2 | 41.2 | 47.5 | 32.4 | 42.7 | |
| OA-DG+SNFTraining Source=F&B, Augmentation-free=false2026.01 | 44.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 40.8 | 54.2 | 56.6 | 40.9 | 40.9 | 51.8 | 26.8 | 41.5 | |
| OA-DGBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Loss function=OA-Loss2023.12 | 43.4 | 8.2 | 10.6 | 8.4 | 24.6 | 20.5 | 22.3 | 4.8 | 6.1 | 25 | 38.4 | 39.7 | 32.8 | 40.2 | 23.8 | 22 | 21.8 | — | — | — | — | — | — | — | — | |
| OA-DGTraining Source=F&B, Augmentation-free=false2026.01 | 43.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 39.5 | 53.1 | 55.9 | 40.4 | 40.6 | 51 | 25.6 | 41.2 | |
| SupConBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Loss function=SupCon2023.12 | 43.2 | 7 | 9.5 | 7.4 | 22.6 | 20.2 | 22.3 | 4.3 | 5.3 | 23 | 37.3 | 38.9 | 31.6 | 40.1 | 24 | 20.1 | 20.9 | — | — | — | — | — | — | — | — | |
| FSCEBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Loss function=FSCE2023.12 | 43.1 | 7.4 | 10.2 | 8.2 | 23.3 | 20.3 | 21.5 | 4.8 | 5.6 | 23.6 | 37.1 | 38 | 31.9 | 40 | 23.2 | 20.4 | 21 | — | — | — | — | — | — | — | — | |
| Photo'Backbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Data augmentation=Photometric distortion2023.12 | 42.7 | 1.6 | 2.7 | 1.9 | 17.9 | 14.1 | 18.7 | 2 | 2.4 | 16.5 | 36 | 39.1 | 27.1 | 39.7 | 18 | 16.4 | 16.9 | — | — | — | — | — | — | — | — | |
| OA-MixBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Data augmentation=OA-Mix2023.12 | 42.7 | 7.2 | 9.6 | 7.7 | 22.8 | 18.8 | 21.9 | 5.4 | 5.2 | 23.6 | 37.3 | 38.7 | 31.9 | 40.2 | 22.2 | 20.2 | 20.8 | — | — | — | — | — | — | — | — | |
| CutoutBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Data augmentation=Cutout2023.12 | 42.5 | 0.6 | 1.2 | 1.2 | 17.8 | 15.9 | 18.9 | 2 | 2.5 | 13.6 | 29.8 | 32.3 | 24.6 | 40.1 | 18.9 | 15.6 | 15.7 | — | — | — | — | — | — | — | — | |
| AutoAug-detBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Data augmentation=AutoAug-det2023.12 | 42.4 | 0.9 | 1.6 | 0.9 | 16.8 | 14.4 | 18.9 | 2 | 1.9 | 16 | 32.9 | 35.2 | 26.3 | 39.4 | 17.9 | 11.6 | 15.8 | — | — | — | — | — | — | — | — | |
| GB-DALTraining Source=F&B, Augmentation-free=true2026.01 | 42.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 39.2 | 50.1 | 53.8 | 36.3 | 39.8 | 48 | 30.7 | 40.3 | |
| StandardBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Training domain=Clean2023.12 | 42.2 | 0.5 | 1.1 | 1.1 | 17.2 | 16.5 | 18.3 | 2.1 | 2.2 | 12.3 | 29.8 | 32 | 24.1 | 40.1 | 18.7 | 15.1 | 15.4 | — | — | — | — | — | — | — | — | |
| NP+SNFTraining Source=F&B, Augmentation-free=false2026.01 | 41.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 41.4 | 55.5 | 55.7 | 39.9 | 38.1 | 47.4 | 14.9 | 38 | |
| FSDR+SNFTraining Source=F&B, Augmentation-free=false2026.01 | 40.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 39 | 55.5 | 55 | 38.8 | 38.7 | 48.4 | 12.2 | 38.7 | |
| AugMixBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Data augmentation=AugMix2023.12 | 39.5 | 5 | 6.8 | 5.1 | 18.3 | 18.1 | 19.3 | 6.2 | 5 | 20.5 | 31.2 | 33.7 | 25.6 | 37.4 | 20.3 | 19.6 | 18.1 | — | — | — | — | — | — | — | — | |
| FACT+SNFTraining Source=F&B, Augmentation-free=false2026.01 | 39.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 38 | 54.4 | 56.9 | 36.1 | 38.4 | 46.8 | 6.9 | 37.5 | |
| DivAlign w/o DivTraining Source=F&B, Augmentation-free=true2026.01 | 38 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 35.6 | 47.4 | 48.5 | 32.8 | 35.2 | 41.5 | 25.4 | 37.1 | |
| ERMTraining Source=F&B, Augmentation-free=true2026.01 | 37.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 37.1 | 45.2 | 53.6 | 32.7 | 37.7 | 47.8 | 5.4 | 38.2 | |
| FSDRTraining Source=F&B, Augmentation-free=false2026.01 | 36.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 36 | 47 | 54.9 | 34.9 | 37.1 | 46.8 | 5 | 33.1 | |
| StylizedBackbone=ResNet-50, Neck=FPN, Detector=Faster R-CNN, Data augmentation=Stylized2023.12 | 36.3 | 4.8 | 6.8 | 4.3 | 19.5 | 18.7 | 18.5 | 2.7 | 3.5 | 17 | 30.5 | 31.9 | 22.7 | 33.9 | 22.6 | 20.8 | 17.2 | — | — | — | — | — | — | — | — | |
| NPTraining Source=F&B, Augmentation-free=false2026.01 | 35.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 37.5 | 44.9 | 53.3 | 30.2 | 35 | 47 | 5 | 33.6 | |
| DALTraining Source=F&B, Augmentation-free=true2026.01 | 35.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 36.2 | 44 | 51.9 | 30.6 | 34.8 | 44.1 | 19 | 23.9 | |
| FACTTraining Source=F&B, Augmentation-free=false2026.01 | 34.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 36.2 | 41.7 | 53.3 | 30 | 34.5 | 45.7 | 2.8 | 34.3 |