Semantic Segmentation on ACDC Snow
70.9mIoUCoDA
Evaluation Results
| Method | Links | |
|---|---|---|
| CoDABackbone=SegFormer, Scene Specificity=Scene-Agnostic2024.03 | 70.9 | |
| DINOv2 + CDTrain. data=ACDC Snow2026.02 | 70.1 | |
| HRDABackbone=SegFormer, Scene Specificity=Scene-Agnostic2024.03 | 69.5 | |
| DINOv2 + CDTrain. data=CityScapes2026.02 | 66.7 | |
| MICBackbone=SegFormer, Scene Specificity=Scene-Agnostic2024.03 | 66.6 | |
| DINOv2Train. data=ACDC Snow2026.02 | 64.5 | |
| DINOv2Train. data=CityScapes2026.02 | 62.3 | |
| STABackbone=SegFormer, Scene Specificity=Scene-Agnostic2024.03 | 58 | |
| SePiCoBackbone=SegFormer, Scene Specificity=Scene-Agnostic2024.03 | 57.9 | |
| DAFormerBackbone=SegFormer, Scene Specificity=Scene-Agnostic2024.03 | 53.7 | |
| DiffTPT + HisTPTModel Scale=Large2024.10 | 49.1 | |
| TPT + HisTPTModel Scale=Large2024.10 | 48.8 | |
| HisTPTModel Scale=Large2024.10 | 48.5 | |
| CoCoOp+CAKIBase Model=SEEM-Large2026.05 | 48.5 | |
| CoOp+CAKIBase Model=SEEM-Large2026.05 | 48.1 | |
| CoCoOpBase Model=SEEM-Large2026.05 | 47.6 | |
| CoOpBase Model=SEEM-Large2026.05 | 47.4 | |
| MR-DCoTDomain Adaptation=Zero-shot, Source=Cityscapes2026.07 | 47.3 | |
| DiffTPTModel Scale=Large2024.10 | 46.3 | |
| ULDASource Domain=Cityscapes, Prompt=driving in snow, Backbone=CLIP-ResNet-50, Base Model=DeepLabv3+2024.04 | 46 | |
| ULDADomain Adaptation=Zero-shot, Source=Cityscapes2026.07 | 46 | |
| TPTModel Scale=Large2024.10 | 45.7 | |
| SEEMModel Scale=Large2024.10 | 45 | |
| SEEM-Large2026.05 | 45 | |
| PØDASource Domain=Cityscapes, Prompt=driving in snow, Backbone=CLIP-ResNet-50, Base Model=DeepLabv3+2024.04 | 43.9 | |
| PØDAAdaptation setting=0-shot, Prompt=driving in snow2024.10 | 43.9 | |
| PØDASource domain=Cityscapes (CS), Target prompt (P)=driving in snow, Zero-shot domain adaptation=true2024.10 | 43.9 | |
| PØDADomain Adaptation=Zero-shot, Source=Cityscapes2026.07 | 43.9 | |
| PIDAAdaptation setting=1-shot, Target image usage=single target image2024.10 | 43.66 | |
| CLIPstylerSource Domain=Cityscapes, Prompt=driving in snow, Backbone=CLIP-ResNet-50, Base Model=DeepLabv3+2024.04 | 41.09 | |
| CLIPstylerSource domain=Cityscapes (CS), Target prompt (P)=driving in snow, Zero-shot domain adaptation=true2024.10 | 41.09 | |
| CLIPstylerDomain Adaptation=Zero-shot, Source=Cityscapes2026.07 | 41.09 | |
| AdaINAdaptation setting=1-shot, Target image usage=single target image2024.10 | 40.25 | |
| source-onlySource Domain=Cityscapes, Prompt=driving in snow, Backbone=CLIP-ResNet-50, Base Model=DeepLabv3+2024.04 | 39.28 | |
| source-only2024.10 | 39.28 | |
| source-onlySource domain=Cityscapes (CS), Target prompt (P)=driving in snow, Zero-shot domain adaptation=true2024.10 | 39.28 | |
| Source-onlyDomain Adaptation=Zero-shot, Source=Cityscapes2026.07 | 39.28 | |
| CoCoOp+CAKIBase Model=SEEM-Tiny2026.05 | 38.5 | |
| CoOp+CAKIBase Model=SEEM-Tiny2026.05 | 38.4 | |
| CoCoOpBase Model=SEEM-Tiny2026.05 | 38.2 | |
| DiffTPT + HisTPTModel Scale=Tiny2024.10 | 38.1 | |
| TPT + HisTPTModel Scale=Tiny2024.10 | 37.8 | |
| HisTPTModel Scale=Tiny2024.10 | 37.1 | |
| CoOpBase Model=SEEM-Tiny2026.05 | 37.1 | |
| DiffTPTModel Scale=Tiny2024.10 | 36.6 | |
| TPTModel Scale=Tiny2024.10 | 36.5 | |
| SEEMModel Scale=Tiny2024.10 | 35.8 | |
| SEEM-Tiny2026.05 | 35.8 |