Semantic Segmentation on Cityscapes v1 (test)
76.5mIoUSwiftNetRN-18 ens+
Evaluation results
| Method | Links | ||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SwiftNetRN-18 ens+GPU=GTX 1080Ti, resolution=2048x1024, # params=24.7M, ImageNet pre-training=true, model_type=ensemble2019.03 | 76.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 18.4 | 18.4 | 218 | 109 | |
| HQSegmentation Model=RefineNet-lw2025.01 | 75.66 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HQSegmentation Model=DeepLabv3+2025.01 | 75.64 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SwiftNetRN-18+GPU=GTX 1080Ti, resolution=2048x1024, # params=11.8M, ImageNet pre-training=true2019.03 | 75.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 39.9 | 39.3 | 104 | 52 | |
| SwiftNetRN-18 pyr+GPU=GTX 1080Ti, resolution=2048x1024, # params=12.9M, ImageNet pre-training=true, model_type=pyramid fusion2019.03 | 75.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 34 | 34 | 114 | 57 | |
| GUNet+GPU=TitanXP, resolution=1024x5122019.03 | 70.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 37.3 | 33.3 | — | — | |
| ERFNet+GPU=TitanX M, resolution=1024x512, # params=20M2019.03 | 69.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 11.2 | 18.4 | 27.7 | 55.4 | |
| ICNetGPU=TitanX M, resolution=2048x10242019.03 | 69.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 30.3 | 49.7 | — | — | |
| UV-M3TL2026.02 | 68.59 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 96.17 | — | — | — | — | |
| ERFNetGPU=TitanX M, resolution=1024x512, # params=20M2019.03 | 68 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 11.2 | 18.4 | 27.7 | 55.4 | |
| Oracle - Target SuperArchitecture=DRN-262017.11 | 67.4 | 97.3 | 79.8 | 88.6 | 32.5 | 48.2 | 56.3 | 63.6 | 73.3 | 89 | 58.9 | 93 | 78.2 | 55.2 | 92.2 | 45 | 67.3 | 39.6 | 49.9 | 73.6 | 89.6 | 94.3 | — | — | — | — | |
| XTasC-Net-ResNet34Backbone=ResNet-342026.02 | 66.51 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 93.56 | — | — | — | — | |
| UniRestoreSegmentation Model=DeepLabv3+2025.01 | 66.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniRestoreSegmentation Model=RefineNet-lw2025.01 | 65.73 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UMT-Net2026.02 | 62.34 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 95.18 | — | — | — | — | |
| AdaMT-Net2026.02 | 61.91 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 94.01 | — | — | — | — | |
| Oracle - Target SuperArchitecture=VGG16-FCN8s2017.11 | 60.3 | 96.4 | 74.5 | 87.1 | 35.3 | 37.8 | 36.4 | 46.9 | 60.1 | 89 | 54.3 | 89.8 | 65.6 | 35.9 | 89.4 | 38.6 | 64.1 | 38.6 | 40.5 | 65.1 | 87.6 | 93.1 | — | — | — | — | |
| ESPNetGPU=TitanX, resolution=1024x512, # params=0.4M2019.03 | 60.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 112 | 108.7 | — | — | |
| NAFNetSegmentation Model=DeepLabv3+2025.01 | 58.41 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NAFNetSegmentation Model=RefineNet-lw2025.01 | 58.19 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PromptIRSegmentation Model=DeepLabv3+2025.01 | 58.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DIPSegmentation Model=RefineNet-lw2025.01 | 57.67 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Denseuncertainty weighting=true2026.02 | 57.57 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 93.67 | — | — | — | — | |
| S3DMT-teacher2026.02 | 57.56 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 92.61 | — | — | — | — | |
| PromptIRSegmentation Model=RefineNet-lw2025.01 | 57.54 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DIPSegmentation Model=DeepLabv3+2025.01 | 57.17 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MTANuncertainty weighting=true2026.02 | 56.55 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 93.56 | — | — | — | — | |
| Split-Wideuncertainty weighting=true2026.02 | 56.36 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 93.08 | — | — | — | — | |
| Cross-Stitchuncertainty weighting=true2026.02 | 56.29 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 93.08 | — | — | — | — | |
| ASAMEvaluation Protocol=Zero-shot, Prompt Type=Box prompt2024.05 | 56 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| URIESegmentation Model=DeepLabv3+2025.01 | 55.88 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PromptIR*Segmentation Model=DeepLabv3+2025.01 | 54.67 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAMEvaluation Protocol=Zero-shot, Prompt Type=Box prompt2024.05 | 54.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MTAL2026.02 | 54.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 91.61 | — | — | — | — | |
| Split-Deepuncertainty weighting=true2026.02 | 54.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 91.26 | — | — | — | — | |
| MLwSGSU2026.02 | 53.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 90.96 | — | — | — | — | |
| DiffBIRSegmentation Model=RefineNet-lw2025.01 | 53.68 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PCGrad2026.02 | 53.59 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 91.45 | — | — | — | — | |
| NAFNet*Segmentation Model=RefineNet-lw2025.01 | 53.29 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MTANuncertainty weighting=false2026.02 | 53.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 91.11 | — | — | — | — | |
| KD4MTL2026.02 | 52.71 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 91.54 | — | — | — | — | |
| DiffBIRSegmentation Model=DeepLabv3+2025.01 | 52.49 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PromptIR*Segmentation Model=RefineNet-lw2025.01 | 52.25 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NAFNet*Segmentation Model=DeepLabv3+2025.01 | 51.91 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Denseuncertainty weighting=false2026.02 | 51.91 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 90.89 | — | — | — | — | |
| DIP*Segmentation Model=DeepLabv3+2025.01 | 51.81 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DiffUIRSegmentation Model=RefineNet-lw2025.01 | 51.46 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| URIESegmentation Model=RefineNet-lw2025.01 | 51.45 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DiffUIRSegmentation Model=DeepLabv3+2025.01 | 51.28 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| URIE*Segmentation Model=DeepLabv3+2025.01 | 50.56 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DIP*Segmentation Model=RefineNet-lw2025.01 | 50.35 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LSSA2026.02 | 50.25 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 83.5 | — | — | — | — | |
| Split-Wideuncertainty weighting=false2026.02 | 50.17 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 90.63 | — | — | — | — | |
| Cross-Stitchuncertainty weighting=false2026.02 | 50.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 90.33 | — | — | — | — | |
| Split-Deepuncertainty weighting=false2026.02 | 49.85 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 88.69 | — | — | — | — | |
| DiffBIR*Segmentation Model=DeepLabv3+2025.01 | 48.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DiffBIR*Segmentation Model=RefineNet-lw2025.01 | 48.56 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| URIE*Segmentation Model=RefineNet-lw2025.01 | 48.23 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DiffUIR*Segmentation Model=DeepLabv3+2025.01 | 47.92 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DiffUIR*Segmentation Model=RefineNet-lw2025.01 | 45.01 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAM + DatasetDMEvaluation Protocol=Zero-shot, Prompt Type=Box prompt2024.05 | 44.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LQSegmentation Model=RefineNet-lw2025.01 | 40.75 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LQSegmentation Model=DeepLabv3+2025.01 | 40.36 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CyCADA pixel+featArchitecture=DRN-262017.11 | 39.5 | 79.1 | 33.1 | 77.9 | 23.4 | 17.3 | 32.1 | 33.3 | 31.8 | 81.5 | 26.7 | 69 | 62.8 | 14.7 | 74.5 | 20.9 | 25.6 | 6.9 | 18.8 | 20.4 | 72.4 | 82.3 | — | — | — | — | |
| CyCADA pixel-onlyArchitecture=DRN-262017.11 | 37 | 63.7 | 24.7 | 69.3 | 21.2 | 17 | 30.3 | 33 | 32 | 80.5 | 25.3 | 62.3 | 62 | 15.1 | 73.1 | 19.8 | 23.6 | 5.5 | 16.2 | 28.7 | 63.8 | 75.4 | — | — | — | — | |
| CyCADA pixel+featArchitecture=VGG16-FCN8s2017.11 | 35.4 | 85.2 | 37.2 | 76.5 | 21.8 | 15 | 23.8 | 22.9 | 21.5 | 80.5 | 31.3 | 60.7 | 50.5 | 9 | 76.9 | 17.1 | 28.2 | 4.5 | 9.8 | 0 | 73.8 | 83.6 | — | — | — | — | |
| CyCADA pixel-onlyArchitecture=VGG16-FCN8s2017.11 | 34.8 | 83.5 | 38.3 | 76.4 | 20.6 | 16.5 | 22.2 | 26.2 | 21.9 | 80.4 | 28.7 | 65.7 | 49.4 | 4.2 | 74.6 | 16 | 26.6 | 2 | 8 | 0 | 73.1 | 82.8 | — | — | — | — | |
| SAM + PGD TuningEvaluation Protocol=Zero-shot, Prompt Type=Box prompt2024.05 | 33.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CyCADA feat-onlyArchitecture=DRN-262017.11 | 31.7 | 78.1 | 31.1 | 71.2 | 10.3 | 14.1 | 29.8 | 28.1 | 20.9 | 74 | 16.8 | 51.9 | 53.6 | 6.1 | 65.4 | 8.2 | 20.9 | 1.8 | 13.9 | 5.9 | 67.4 | 78.4 | — | — | — | — | |
| SAM + DAT TuningEvaluation Protocol=Zero-shot, Prompt Type=Box prompt2024.05 | 31.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CyCADA feat-onlyArchitecture=VGG16-FCN8s2017.11 | 29.2 | 85.6 | 30.7 | 74.7 | 14.4 | 13 | 17.6 | 13.7 | 5.8 | 74.6 | 15.8 | 69.9 | 38.2 | 3.5 | 72.3 | 16 | 5 | 0.1 | 3.6 | 0 | 71.5 | 82.5 | — | — | — | — | |
| FCNs in the wildArchitecture=VGG16-FCN8s2017.11 | 27.1 | 70.4 | 32.4 | 62.1 | 14.9 | 5.4 | 10.9 | 14.2 | 2.7 | 79.2 | 21.3 | 64.6 | 44.1 | 4.2 | 70.4 | 8 | 7.3 | 0 | 3.5 | 0 | — | — | — | — | — | — | |
| Source onlyArchitecture=DRN-262017.11 | 21.7 | 42.7 | 26.3 | 51.7 | 5.5 | 6.8 | 13.8 | 23.6 | 6.9 | 75.5 | 11.5 | 36.8 | 49.3 | 0.9 | 46.7 | 3.4 | 5 | 0 | 5 | 1.4 | 47.4 | 62.5 | — | — | — | — | |
| Source onlyArchitecture=VGG16-FCN8s2017.11 | 17.9 | 26 | 14.9 | 65.1 | 5.5 | 12.9 | 8.9 | 6 | 2.5 | 70 | 2.9 | 47 | 24.5 | 0 | 40 | 12.1 | 1.5 | 0 | 0 | 0 | 41.9 | 54 | — | — | — | — |