Semantic Segmentation on Dark Zurich (test)
63.9mIoURefign-HRDA
Evaluation results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Refign-HRDASource Domain=Cityscapes, Input Protocol=Non-standard resizing protocol2022.07 | 63.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Refign-DAFormerArchitecture=SegFormer2022.07 | 56.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HRDASource Domain=Cityscapes, Input Protocol=Non-standard resizing protocol2022.07 | 55.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SePiCo (DistCL)Backbone=MiT-B52022.04 | 54.2 | 93.2 | 68.1 | 73.7 | 32.8 | 16.3 | 54.6 | 49.5 | 48.1 | 74.2 | 31 | 86.3 | 57.9 | 50.9 | 82.4 | 52.2 | 1.3 | 83.8 | 43.9 | 29.8 | |
| SePiCo (DAFormer)Source Domain=Cityscapes, Input Protocol=Non-standard resizing protocol2022.07 | 54.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DAFormerArchitecture=SegFormer2022.07 | 53.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMAArchitecture=SegFormer2023.03 | 53.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SePiCo (BankCL)Backbone=MiT-B52022.04 | 53.3 | 91.1 | 61.2 | 73.4 | 31.9 | 18 | 51.6 | 48.6 | 47.7 | 72.8 | 33 | 85.5 | 57 | 51.1 | 80.6 | 48.4 | 3.1 | 84.6 | 45.3 | 28.2 | |
| SePiCo (ProtoCL)Backbone=MiT-B52022.04 | 52.7 | 90.1 | 57.7 | 75 | 34.9 | 16.4 | 53.5 | 47 | 47.8 | 70.1 | 31.7 | 84.1 | 57.3 | 53.3 | 80.5 | 42.4 | 2.3 | 83.6 | 42.6 | 30.1 | |
| URMA + SimTArchitecture=SegFormer2023.03 | 50.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| URMAArchitecture=SegFormer2023.03 | 49.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DAFormerBackbone=MiT-B52022.04 | 48.5 | 92 | 63 | 67.2 | 28.9 | 13.1 | 44 | 42 | 42.3 | 70.7 | 28.2 | 83.6 | 51.1 | 39.1 | 76.4 | 31.7 | 0 | 78.3 | 43.9 | 26.5 | |
| CCDistillModel=RefineNet, Backbone=ResNet-1012022.05 | 47.5 | 89.6 | 58.1 | 70.6 | 36.6 | 22.5 | 33 | 27 | 30.5 | 68.3 | 33 | 80.9 | 42.3 | 40.1 | 69.4 | 58.1 | 0.1 | 72.6 | 47.7 | 21.3 | |
| CCDistillArchitecture=RefineNet2022.07 | 47.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DANIAArchitecture=PSPNet2022.07 | 47 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DANNetArchitecture=PSPNet2022.07 | 45.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CDAdaArchitecture=RefineNet2022.07 | 45 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DANNetBase Model=RefineNet2022.05 | 44.3 | 90 | 54 | 74.8 | 41 | 21.1 | 25 | 26.8 | 30.2 | 72 | 26.2 | 84 | 47 | 33.9 | 68.2 | 19 | 0.3 | 66.4 | 38.3 | 23.6 | |
| TENTArchitecture=SegFormer2023.03 | 42.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HCLArchitecture=SegFormer2023.03 | 42.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MGCDABackbone=ResNet-1012020.05 | 42.5 | 80.3 | 49.3 | 66.2 | 7.8 | 11 | 41.4 | 38.9 | 39 | 64.1 | 18 | 55.8 | 52.1 | 53.5 | 74.7 | 66 | 0 | 37.5 | 29.1 | 22.7 | |
| MGCDATraining Data Source=source and target data2021.08 | 42.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MGCDA2022.05 | 42.5 | 80.3 | 49.3 | 66.2 | 7.8 | 11 | 41.4 | 38.9 | 39 | 64.1 | 18 | 55.8 | 52.1 | 53.5 | 74.7 | 66 | 0 | 37.5 | 29.1 | 22.7 | |
| MGCDAArchitecture=RefineNet2022.07 | 42.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GCMA2019.01 | 42 | 81.7 | 46.9 | 58.8 | 22 | 20 | 41.2 | 40.5 | 41.6 | 64.8 | 31 | 32.1 | 53.5 | 47.5 | 75.5 | 39.2 | 0 | 49.6 | 30.7 | 21 | |
| GCMABackbone=ResNet-1012020.05 | 42 | 81.7 | 46.9 | 58.8 | 22 | 20 | 41.2 | 40.5 | 41.6 | 64.8 | 31 | 32.1 | 53.5 | 47.5 | 75.5 | 39.2 | 0 | 49.6 | 30.7 | 21 | |
| GCMATraining Data Source=source and target data2021.08 | 42 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GCMA2022.05 | 42 | 81.7 | 46.9 | 58.8 | 22 | 20 | 41.2 | 40.5 | 41.6 | 64.8 | 31 | 32.1 | 53.5 | 47.5 | 75.5 | 39.2 | 0 | 49.6 | 30.7 | 21 | |
| GCMAArchitecture=RefineNet2022.07 | 42 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Source modelArchitecture=SegFormer2023.03 | 41.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Refign-DACSArchitecture=DeepLabv22022.07 | 41.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DACSArchitecture=DeepLabv22022.07 | 36.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SegF. MiT-B5Backbone=MiT-B5, Number of Parameters=84.7M2022.04 | 35 | 80.3 | 37.1 | 57.5 | 28.1 | 7.9 | 35.5 | 33.2 | 29.3 | 41.7 | 14.8 | 4.7 | 48.9 | 48 | 66.6 | 5.7 | 7.9 | 63.3 | 31.4 | 23.3 | |
| W-RefineNetTraining Data Source=source data only2021.08 | 34.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UDAclusteringTraining Set=Cityscapes, Domain Adaptation Target=DZ-night2022.05 | 33.3 | 85.5 | 40.9 | 59.2 | 31.2 | 19.5 | 24 | 29.9 | 29.4 | 30.6 | 11.2 | 18.4 | 39.1 | 49.7 | 61.5 | 34.9 | 0 | 25.8 | 23.2 | 19 | |
| DMAda2019.01 | 32.1 | 75.5 | 29.1 | 48.6 | 21.3 | 14.3 | 34.3 | 36.8 | 29.9 | 49.4 | 13.8 | 0.4 | 43.3 | 50.2 | 69.4 | 18.4 | 0 | 27.6 | 34.9 | 11.9 | |
| DMAdaBackbone=ResNet-1012020.05 | 32.1 | 75.5 | 29.1 | 48.6 | 21.3 | 14.3 | 34.3 | 36.8 | 29.9 | 49.4 | 13.8 | 0.4 | 43.3 | 50.2 | 69.4 | 18.4 | 0 | 27.6 | 34.9 | 11.9 | |
| DMAdaTraining Data Source=source and target data2021.08 | 32.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DMAda2022.05 | 32.1 | 75.5 | 29.1 | 48.6 | 21.3 | 14.3 | 34.3 | 36.8 | 29.9 | 49.4 | 13.8 | 0.4 | 43.3 | 50.2 | 69.4 | 18.4 | 0 | 27.6 | 34.9 | 11.9 | |
| DMAdaArchitecture=RefineNet2022.07 | 32.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RefineNet-AdaBNTraining Data Source=source data only2021.08 | 31.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BDLBackbone=ResNet-101, Adaptation Protocol=Cityscapes→DZ-night2020.05 | 30.8 | 85.3 | 41.1 | 61.9 | 32.7 | 17.4 | 20.6 | 11.4 | 21.3 | 29.4 | 8.9 | 1.1 | 37.4 | 22.1 | 63.2 | 28.2 | 0 | 47.7 | 39.4 | 15.7 | |
| BDLTraining Data Source=source and target data2021.08 | 30.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BDLTraining Set=Cityscapes, Domain Adaptation Target=DZ-night2022.05 | 30.8 | 85.3 | 41.1 | 61.9 | 32.7 | 17.4 | 20.6 | 11.4 | 21.3 | 29.4 | 8.9 | 1.1 | 37.4 | 22.1 | 63.2 | 28.2 | 0 | 47.7 | 39.4 | 15.7 | |
| RefineNetTraining Data Source=source data only2021.08 | 30.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdaptSegNetadaptation_source=Cityscapes, adaptation_target=DZ-night2019.01 | 30.4 | 86.1 | 44.2 | 55.1 | 22.2 | 4.8 | 21.1 | 5.6 | 16.7 | 37.2 | 8.4 | 1.2 | 35.9 | 26.7 | 68.2 | 45.1 | 0 | 50.1 | 33.9 | 15.6 | |
| AdaptSegNetBackbone=ResNet-101, Adaptation Protocol=Cityscapes→DZ-night2020.05 | 30.4 | 86.1 | 44.2 | 55.1 | 22.2 | 4.8 | 21.1 | 5.6 | 16.7 | 37.2 | 8.4 | 1.2 | 35.9 | 26.7 | 68.2 | 45.1 | 0 | 50.1 | 33.9 | 15.6 | |
| AdaptSegNetTraining Data Source=source and target data2021.08 | 30.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdaptSegNetTraining Set=Cityscapes, Domain Adaptation Target=DZ-night2022.05 | 30.4 | 86.1 | 44.2 | 55.1 | 22.2 | 4.8 | 21.1 | 5.6 | 16.7 | 37.2 | 8.4 | 1.2 | 35.9 | 26.7 | 68.2 | 45.1 | 0 | 50.1 | 33.9 | 15.6 | |
| ADVENTBackbone=ResNet-101, Adaptation Protocol=Cityscapes→DZ-night2020.05 | 29.7 | 85.8 | 37.9 | 55.5 | 27.7 | 14.5 | 23.1 | 14 | 21.1 | 32.1 | 8.7 | 2 | 39.9 | 16.6 | 64 | 13.8 | 0 | 58.8 | 28.5 | 20.7 | |
| ADVENTTraining Data Source=source and target data2021.08 | 29.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ADVENTTraining Set=Cityscapes, Domain Adaptation Target=DZ-night2022.05 | 29.7 | 85.8 | 37.9 | 55.5 | 27.7 | 14.5 | 23.1 | 14 | 21.1 | 32.1 | 8.7 | 2 | 39.9 | 16.6 | 64 | 13.8 | 0 | 58.8 | 28.5 | 20.7 | |
| AdaptSegNetadaptation_source=Cityscapes2019.01 | 28.8 | 79 | 21.8 | 53 | 13.3 | 11.2 | 22.5 | 20.2 | 22.1 | 43.5 | 10.4 | 18 | 37.4 | 33.8 | 64.1 | 6.4 | 0 | 52.3 | 30.4 | 7.4 | |
| DeepLab-v2Backbone=ResNet-101, Mode=Daytime trained baseline2020.05 | 28.8 | 79 | 21.8 | 53 | 13.3 | 11.2 | 22.5 | 20.2 | 22.1 | 43.5 | 10.4 | 18 | 37.4 | 33.8 | 64.1 | 6.4 | 0 | 52.3 | 30.4 | 7.4 | |
| DeepLab-v2Training Set=Cityscapes2022.05 | 28.8 | 79 | 21.8 | 53 | 13.3 | 11.2 | 22.5 | 20.2 | 22.1 | 43.5 | 10.4 | 18 | 37.4 | 33.8 | 64.1 | 6.4 | 0 | 52.3 | 30.4 | 7.4 | |
| RefineNet2019.01 | 28.5 | 68.8 | 23.2 | 46.8 | 20.8 | 12.6 | 29.8 | 30.4 | 26.9 | 43.1 | 14.3 | 0.3 | 36.9 | 49.7 | 63.6 | 6.8 | 0.2 | 24 | 33.6 | 9.3 | |
| RefineNetBackbone=ResNet-101, Mode=Daytime trained baseline2020.05 | 28.5 | 68.8 | 23.2 | 46.8 | 20.8 | 12.6 | 29.8 | 30.4 | 26.9 | 43.1 | 14.3 | 0.3 | 36.9 | 49.7 | 63.6 | 6.8 | 0.2 | 24 | 33.6 | 9.3 | |
| RefineNetTraining Set=Cityscapes2022.05 | 28.5 | 68.8 | 23.2 | 46.8 | 20.8 | 12.6 | 29.8 | 30.4 | 26.9 | 43.1 | 14.3 | 0.3 | 36.9 | 49.7 | 63.6 | 6.8 | 0.2 | 24 | 33.6 | 9.3 |