Semantic Segmentation on Cityscapes GTA5 source 1.0 (val)
76.1mIoUILM-ASSL
Evaluation results
| Method | Links | |||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ILM-ASSLNet.=DeepLab-v3+, Annotation budget=5%2023.01 | 76.1 | 96.9 | 77.8 | 91.6 | 46.7 | 56 | 63.2 | 70.8 | 77.4 | 54.9 | 94.5 | 82.3 | 61.2 | 94.9 | 79.3 | 65.8 | 77.6 | 91.9 | 88.1 | 75.3 | — | |
| ILM-ASSLNet.=DeepLab-v3+, Annotation budget=2.2%2023.01 | 75 | 96.5 | 75.6 | 91.2 | 46.7 | 53.6 | 62.1 | 70.3 | 76 | 52.1 | 94.1 | 82 | 60.8 | 94.4 | 83.1 | 61.2 | 75.8 | 91.4 | 86.4 | 71.9 | — | |
| ILM-ASSLNet.=DeepLab-v2, Annotation budget=2.2%2023.01 | 72.5 | 96.4 | 74.6 | 91.1 | 45.9 | 52.4 | 59.4 | 67.9 | 68.3 | 50 | 92.8 | 76.2 | 57.2 | 93.6 | 78.2 | 58.4 | 72.1 | 91.4 | 81.3 | 69.5 | — | |
| RIPUNet.=DeepLab-v3+, Annotation budget=5%2023.01 | 71.2 | 97 | 77.3 | 90.4 | 54.6 | 53.2 | 47.7 | 55.9 | 64.1 | 59.2 | 93.2 | 75 | 54.8 | 92.7 | 73 | 55.5 | 70.3 | 90.2 | 79.7 | 68.9 | — | |
| ILM-ASSLNet.=DeepLab-v3+, Annotation budget=1%2023.01 | 70 | 95.2 | 67 | 90.9 | 47.4 | 49.6 | 60.9 | 68.2 | 67.5 | 44.6 | 91.5 | 81.3 | 60.5 | 93.9 | 67.2 | 54.7 | 74.8 | 90.9 | 76.6 | 47.9 | — | |
| RIPUNet.=DeepLab-v2, Annotation budget=2.2%2023.01 | 69.6 | 96.5 | 74.1 | 89.7 | 53.1 | 51 | 43.8 | 53.4 | 62.2 | 57.6 | 92.6 | 73 | 53 | 92.8 | 73.8 | 55.6 | 70 | 90 | 78.5 | 62 | — | |
| LabORNet.=DeepLab-v2, Annotation budget=2.2%2023.01 | 66.6 | 96.6 | 77 | 89.6 | 47.8 | 50.7 | 48 | 56.6 | 63.5 | 57.8 | 91.6 | 72 | 47.3 | 91.7 | 62.1 | 47.9 | 65.3 | 89.5 | 61.9 | 48.9 | — | |
| MADANet.=DeepLab-v3+, Annotation budget=5%2023.01 | 64.9 | 95.1 | 69.8 | 88.5 | 43.3 | 48.7 | 45.7 | 53.3 | 59.2 | 46.7 | 91.5 | 73.9 | 50.1 | 91.2 | 60.6 | 51.6 | 68.7 | 89.1 | 56.9 | 48.4 | — | |
| DAP+ProDANet.=DeepLab-v2, Annotation budget=0%2023.01 | 59.8 | 94.5 | 63.1 | 89.1 | 29.8 | 47.5 | 50.4 | 56.7 | 58.7 | 50.2 | 87 | 73.6 | 38.6 | 91.3 | 50.2 | 50.2 | 63.5 | 89.5 | 52.9 | 0 | — | |
| AADANet.=DeepLab-v3+, Annotation budget=5%2023.01 | 59.3 | 92.2 | 59.9 | 87.3 | 36.4 | 45.7 | 46.1 | 50.6 | 59.5 | 44 | 90.2 | 69.7 | 38.2 | 90 | 55.3 | 32.6 | 62.9 | 88.3 | 45.1 | 32 | — | |
| ProDA+CRABackbone=ResNet-1012021.09 | 58.6 | 89.4 | 60 | 81 | 49.2 | 44.8 | 45.5 | 53.6 | 55 | 51.9 | 85.6 | 72.3 | 40.8 | 88.5 | 44.3 | 51.7 | 57.9 | 89.4 | 53.4 | 0 | — | |
| ProDABackbone=ResNet-1012021.09 | 57.5 | 87.8 | 56 | 79.7 | 46.3 | 44.8 | 45.6 | 53.5 | 53.5 | 45.2 | 82.1 | 70.7 | 39.2 | 88.8 | 45.5 | 48.9 | 56.4 | 88.6 | 59.4 | 1 | — | |
| ProDANet.=DeepLab-v2, Annotation budget=0%2023.01 | 57.5 | 87.8 | 56 | 79.7 | 46.3 | 44.8 | 45.6 | 53.5 | 53.5 | 45.2 | 82.1 | 70.7 | 39.2 | 88.8 | 45.5 | 48.9 | 56.4 | 88.6 | 59.4 | 1 | — | |
| IAST+CRABackbone=ResNet-1012021.09 | 54.1 | 93.3 | 57.8 | 87.2 | 42.1 | 29.6 | 40.9 | 50.8 | 51 | 28.2 | 87.7 | 67.8 | 34.4 | 87.5 | 29.8 | 40.8 | 52.7 | 86 | 46.3 | 14.2 | — | |
| SACNet.=DeepLab-v2, Annotation budget=0%2023.01 | 53.8 | 90.4 | 53.9 | 86.6 | 42.4 | 27.3 | 45.1 | 48.5 | 42.7 | 40.1 | 86.1 | 67.5 | 29.7 | 88.5 | 49.1 | 26.6 | 45.3 | 87.4 | 54.6 | 9.8 | — | |
| FADA+CRABackbone=ResNet-1012021.09 | 52.2 | 91.6 | 53.6 | 85.5 | 42.6 | 18.7 | 34.8 | 36.3 | 18.9 | 45 | 89 | 66.2 | 39 | 87.6 | 42.3 | 42 | 44.6 | 87.8 | 51.3 | 14.9 | — | |
| IASTBackbone=ResNet-1012021.09 | 52.2 | 94.1 | 58.8 | 85.4 | 39.7 | 29.2 | 25.1 | 43.1 | 34.2 | 34.6 | 88.7 | 62.7 | 30.3 | 87.6 | 42.3 | 35.2 | 40.2 | 84.8 | 50.3 | 24.7 | — | |
| MetaCorrectionmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 52.1 | 92.8 | 58.1 | 86.2 | 39.7 | 33.1 | 36.3 | 42 | 38.6 | 85.5 | 37.8 | 87.6 | 62.8 | 31.7 | 84.8 | 35.7 | 50.3 | — | — | — | — | |
| DMLCmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 51.2 | 92.5 | 55.1 | 85.9 | 36.9 | 32.4 | 34.7 | 41.4 | 37 | 85.3 | 37.8 | 87.4 | 62.7 | 31.8 | 84.5 | 36.8 | 48.2 | — | — | — | — | |
| LTIRmechanism=Adversarial Learning (AL), backbone=ResNet-1012021.03 | 50.2 | 92.9 | 55 | 85.3 | 34.2 | 31.1 | 34.9 | 40.7 | 34 | 85.2 | 40.1 | 87.1 | 61 | 31.1 | 82.5 | 32.3 | 42.9 | — | — | — | — | |
| CAG-UDAmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 50.2 | 90.4 | 51.6 | 83.8 | 34.2 | 27.8 | 38.4 | 25.3 | 48.4 | 85.4 | 38.2 | 78.1 | 58.6 | 34.6 | 84.7 | 21.9 | 42.7 | — | — | — | — | |
| FADABackbone=ResNet-1012021.09 | 50.1 | 91 | 50.6 | 86 | 43.4 | 29.8 | 36.8 | 43.4 | 25 | 38.3 | 87.4 | 64 | 38 | 85.2 | 31.6 | 25.4 | 37.1 | 86.8 | 46.1 | 6.5 | — | |
| ASSNet.=DeepLab-v2, Annotation budget=0%2023.01 | 49.2 | 90.6 | 44.7 | 84.8 | 34.3 | 28.7 | 31.6 | 35 | 37.6 | 43.3 | 85.3 | 57 | 31.5 | 83.8 | 42.6 | 30.4 | 39 | 84.7 | 48.5 | 1.9 | — | |
| MLSLmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 49 | 89 | 45.2 | 78.2 | 22.9 | 27.3 | 37.4 | 46.1 | 43.8 | 82.9 | 18.6 | 61.2 | 60.4 | 26.7 | 85.4 | 35.9 | 44.9 | — | — | — | — | |
| PyCDAmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 47.4 | 90.5 | 36.3 | 84.4 | 32.4 | 28.7 | 34.6 | 36.4 | 31.5 | 86.8 | 37.9 | 78.5 | 62.3 | 21.5 | 85.6 | 27.9 | 34.8 | — | — | — | — | |
| FADA+CRABackbone=VGG-162021.09 | 47.2 | 93.3 | 59.3 | 84.6 | 24.8 | 26.1 | 36.2 | 33.5 | 30.2 | 38.4 | 85.6 | 60.4 | 29.6 | 85 | 24.7 | 24.3 | 17.5 | 84.2 | 39 | 20.4 | — | |
| CRSTmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 47.1 | 91 | 55.4 | 80 | 33.7 | 21.4 | 37.3 | 32.9 | 24.5 | 85 | 34.1 | 80.8 | 57.7 | 24.6 | 84.1 | 27.8 | 30.1 | — | — | — | — | |
| MRKLDNet.=DeepLab-v2, Annotation budget=0%2023.01 | 47.1 | 91 | 55.4 | 80 | 33.7 | 21.4 | 37.3 | 32.9 | 24.5 | 34.1 | 80.8 | 57.7 | 24.6 | 84.1 | 27.8 | 26 | 42.3 | 85 | 30.1 | 26.9 | — | |
| ADVENT+CRABackbone=ResNet-1012021.09 | 46.7 | 90 | 39.9 | 83.5 | 33.3 | 27.5 | 28.7 | 35 | 27.6 | 37.2 | 80 | 57.9 | 29.4 | 85.1 | 39.3 | 26.4 | 37.6 | 85.2 | 44.2 | 0 | — | |
| PatchAlignmechanism=Adversarial Learning (AL), backbone=ResNet-1012021.03 | 46.5 | 92.3 | 51.9 | 82.1 | 29.2 | 25.1 | 24.5 | 33.8 | 33 | 82.4 | 32.8 | 82.2 | 58.6 | 27.2 | 84.3 | 33.4 | 46.3 | — | — | — | — | |
| MaxSquaremechanism=Self-training (ST), backbone=ResNet-1012021.03 | 46.4 | 89.4 | 43 | 82.1 | 30.5 | 21.3 | 30.3 | 34.7 | 24 | 85.3 | 39.4 | 78.2 | 63 | 22.9 | 84.6 | 36.4 | 43 | — | — | — | — | |
| IntraDAmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 46.3 | 90.6 | 37.1 | 82.6 | 30.1 | 19.1 | 29.5 | 32.4 | 20.6 | 85.7 | 40.5 | 79.7 | 58.7 | 31.1 | 86.3 | 31.5 | 48.3 | — | — | — | — | |
| IntraDABackbone=ResNet-1012021.09 | 46.3 | 90.6 | 36.1 | 82.6 | 29.5 | 21.3 | 27.6 | 31.4 | 23.1 | 39.3 | 80.2 | 59.3 | 29.4 | 86.4 | 33.6 | 32.7 | 37.6 | 85.2 | 53.9 | 0 | — | |
| CBSTmechanism=Self-training (ST), backbone=ResNet-1012021.03 | 45.9 | 91.8 | 53.5 | 80.5 | 32.7 | 21 | 34 | 28.9 | 20.4 | 83.9 | 34.2 | 80.9 | 53.1 | 24 | 82.7 | 30.3 | 35.9 | — | — | — | — | |
| CBSTNet.=DeepLab-v2, Annotation budget=0%2023.01 | 45.9 | 91.8 | 53.5 | 80.5 | 32.7 | 21 | 34 | 28.9 | 20.4 | 34.2 | 80.9 | 53.1 | 24 | 82.7 | 30.3 | 25.9 | 42.8 | 83.9 | 35.9 | 16 | — | |
| FADABackbone=VGG-162021.09 | 43.8 | 92.3 | 51.1 | 83.7 | 33.1 | 29.1 | 28.5 | 28 | 21 | 32.6 | 85.3 | 55.2 | 28.8 | 83.5 | 24.4 | 21.1 | 15.2 | 82.6 | 37.4 | 0 | — | |
| ADVENTBackbone=ResNet-1012021.09 | 43.8 | 89.9 | 36.5 | 81.6 | 29.2 | 25.2 | 28.5 | 32.3 | 22.4 | 34 | 77.1 | 57.4 | 27.9 | 83.7 | 29.4 | 28.4 | 23.3 | 83.9 | 39.1 | 1.5 | — | |
| AdaptSegNet+CRABackbone=ResNet-1012021.09 | 43.4 | 87.7 | 33.5 | 81.6 | 33.3 | 23.7 | 22.4 | 26.1 | 16 | 38.5 | 72.7 | 52.7 | 21.2 | 84.2 | 44.9 | 26.2 | 33 | 83.6 | 43 | 0 | — | |
| AdaptSegNetmechanism=Adversarial Learning (AL), backbone=ResNet-1012021.03 | 42.4 | 86.5 | 36 | 79.9 | 23.4 | 23.3 | 23.9 | 35.2 | 14.8 | 83.4 | 33.3 | 75.6 | 58.5 | 27.6 | 73.7 | 32.5 | 35.4 | — | — | — | — | |
| AdaptSegNetBackbone=ResNet-1012021.09 | 42.4 | 86.5 | 36 | 79.9 | 23.4 | 23.3 | 23.9 | 35.2 | 14.8 | 33.3 | 75.6 | 58.5 | 27.6 | 73.7 | 32.5 | 30.1 | 28.1 | 83.4 | 35.4 | 3.9 | — | |
| OursNetwork=VGG-16, SRSS Setting=GTA5 to Cityscapes2021.08 | 37.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.8 | |
| Source OnlyBackbone=ResNet-1012021.09 | 36.8 | 65 | 16.1 | 68.7 | 18.6 | 16.8 | 21.3 | 31.4 | 11.2 | 22 | 78 | 54.4 | 33.8 | 73.9 | 12.7 | 28.1 | 19.7 | 83 | 30.7 | 13.7 | — | |
| Source onlybackbone=ResNet-1012021.03 | 36.6 | 75.8 | 16.8 | 77.2 | 12.5 | 21 | 25.5 | 30.1 | 20.1 | 81.3 | 24.6 | 70.3 | 53.8 | 26.4 | 49.9 | 17.2 | 25.9 | — | — | — | — | |
| Source OnlyNet.=DeepLab-v2, Annotation budget=0%2023.01 | 36.6 | 75.8 | 16.8 | 77.2 | 12.5 | 21 | 25.5 | 30.1 | 20.1 | 24.6 | 70.3 | 53.8 | 26.4 | 49.9 | 17.2 | 25.3 | 36 | 81.3 | 25.9 | 6.5 | — | |
| ADVENTBackbone=VGG-162021.09 | 36.1 | 86.9 | 28.7 | 78.7 | 28.5 | 25.2 | 17.1 | 20.3 | 10.9 | 26.4 | 70.2 | 47.1 | 8.4 | 81.5 | 26 | 11.7 | 1.6 | 80 | 17.2 | 18.9 | — | |
| SWNetwork=VGG-16, SRSS Setting=GTA5 to Cityscapes2021.08 | 35.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdaptSegNetBackbone=VGG-162021.09 | 35 | 87.3 | 29.8 | 78.6 | 21.1 | 18.2 | 22.5 | 21.5 | 11 | 29.6 | 71.3 | 46.8 | 6.5 | 80.1 | 23 | 10.6 | 0.3 | 79.7 | 26.9 | 0 | — | |
| Source OnlyNetwork=VGG-16, SRSS Setting=GTA5 to Cityscapes2021.08 | 31.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Source OnlyBackbone=VGG-162021.09 | 28.3 | 35.4 | 13.2 | 72.1 | 16.7 | 11.6 | 20.7 | 22.5 | 13.1 | 7.6 | 66.1 | 41.1 | 19 | 69.8 | 15.2 | 16.2 | 4.7 | 76 | 16.3 | 0 | — |