Semantic Segmentation on Cityscapes 1/16 (186 labeled samples)
84.1mIoUHVLFormer
Evaluation Results
| Method | Links | |
|---|---|---|
| HVLFormerBB=SigLIP22025.06 | 84.1 | |
| UniMatch V2Venue=Ours, Encoder=DINOV2-B, #Params=97.5M2024.10 | 83.6 | |
| UniMV2BB=D.v2-B2025.06 | 83.6 | |
| UniMatch V2 (published)Encoder=DINOv2-B2026.07 | 83.6 | |
| UniMatchV2-B + PrevMatchBackbone=DINOv2-B, Resolution=756x7562024.05 | 83.4 | |
| PixConEncoder=DINOv2-B2026.07 | 83.2 | |
| UniMatch V2 (our repro)Encoder=DINOv2-B2026.07 | 83.16 | |
| HVLFormerBB=ViT-B2025.06 | 83 | |
| UniMatchV2-BBackbone=DINOv2-B, Resolution=756x7562024.05 | 82.6 | |
| UniMatchV2-S + PrevMatchBackbone=DINOv2-S2024.05 | 81.7 | |
| POSBB=D.v2-S2025.06 | 81.4 | |
| HVLFormerBB=E.02-S2025.06 | 81.3 | |
| Labeled OnlyEncoder=DINOV2-B, #Params=97.5M2024.10 | 80.8 | |
| UniMatchV2-SBackbone=DINOv2-S2024.05 | 80.6 | |
| UniMatch V2Venue=Ours, Encoder=DINOV2-S, #Params=24.8M2024.10 | 80.6 | |
| UniMatch V2 (published)Encoder=DINOv2-S2026.07 | 80.6 | |
| BeyondPixelsVenue=ECCV'24, Encoder=RN-101, #Params=59.5M2024.10 | 78.5 | |
| Pseudo-SDBB=ViT-B2025.06 | 78.5 | |
| BeyondPixelsEncoder=RN-1012026.07 | 78.5 | |
| AllSparkBackbone=SegFormer-B5, Evaluation Protocol=OHEM, Sliding window2024.03 | 78.33 | |
| AllSparkBackbone=Transformer-based2024.05 | 78.3 | |
| AllSparkVenue=CVPR'24, Encoder=MiT-B5, #Params=89.3M2024.10 | 78.3 | |
| SemiVLBackbone=ViT-B/162023.11 | 77.9 | |
| SemiVLBackbone=Transformer-based2024.05 | 77.9 | |
| SemiVLVenue=ECCV'24, Encoder=CLIP-B, #Params=88.0M2024.10 | 77.9 | |
| SemiVLBB=ViT-B2025.06 | 77.9 | |
| SemiVLEncoder=CLIP-B2026.07 | 77.9 | |
| UniMatch + PrevMatchBackbone=ResNet-1012024.05 | 77.7 | |
| CorrMatchBackbone=ResNet-1012024.05 | 77.3 | |
| CorrMatchVenue=CVPR'24, Encoder=RN-101, #Params=59.5M2024.10 | 77.3 | |
| CorrMatchEncoder=RN-1012026.07 | 77.3 | |
| CFCGEvaluation Protocol=OHEM, Sliding window2024.03 | 77.28 | |
| Labeled OnlyEncoder=DINOV2-S, #Params=24.8M2024.10 | 77.2 | |
| DDFPVenue=CVPR'24, Encoder=RN-101, #Params=59.5M2024.10 | 77.1 | |
| LogicDiagEvaluation Protocol=OHEM, Sliding window2024.03 | 76.83 | |
| LogicDiagBackbone=ResNet-1012023.11 | 76.8 | |
| Dual TeacherBackbone=ResNet-1012024.05 | 76.8 | |
| LogicDiagVenue=ICCV'23, Encoder=RN-101, #Params=59.5M2024.10 | 76.8 | |
| UniMatchBackbone=ResNet-1012023.11 | 76.6 | |
| UniMatchBackbone=ViT-B/16, Note=re-produced2023.11 | 76.6 | |
| UniMatchEvaluation Protocol=OHEM, Sliding window2024.03 | 76.6 | |
| AugSeg + PrevMatchBackbone=ResNet-1012024.05 | 76.6 | |
| UniMatchBackbone=ResNet-1012024.05 | 76.6 | |
| UniMatch V1Venue=CVPR'23, Encoder=RN-101, #Params=59.5M2024.10 | 76.6 | |
| UniMatch V1Venue=CVPR'23, Encoder=CLIP-B, #Params=88.0M2024.10 | 76.6 | |
| DAWVenue=NeurIPS'23, Encoder=RN-101, #Params=59.5M2024.10 | 76.6 | |
| UniM V1†BB=ViT-B2025.06 | 76.6 | |
| UniMatchEncoder=RN-1012026.07 | 76.6 | |
| AELBackbone=ResNet-1012023.11 | 75.8 | |
| FPLEvaluation Protocol=OHEM, Sliding window2024.03 | 75.74 | |
| 3-CPSBackbone=ResNet-1012023.11 | 75.7 | |
| Diverse Co-T.Backbone=ResNet-1012024.05 | 75.7 | |
| Diverse CoTVenue=ICCV'23, Encoder=RN-101, #Params=59.5M2024.10 | 75.7 | |
| TQDMBB=SigLIP22025.06 | 75.5 | |
| AugSegEvaluation Protocol=OHEM, Sliding window2024.03 | 75.22 | |
| AugSegBackbone=ResNet-1012024.05 | 75.2 | |
| AugSegVenue=CVPR'23, Encoder=RN-101, #Params=59.5M2024.10 | 75.2 | |
| AugSegEncoder=RN-1012026.07 | 75.2 | |
| ESLEvaluation Protocol=OHEM, Sliding window2024.03 | 75.12 | |
| ESLBackbone=ResNet-1012023.11 | 75.1 | |
| ESLVenue=ICCV'23, Encoder=RN-101, #Params=59.5M2024.10 | 75.1 | |
| CCVCBackbone=ResNet-50, Crop Size=712, Partition Protocol=1/162023.03 | 74.9 | |
| U2PLBackbone=ResNet-1012023.11 | 74.9 | |
| U2PLEvaluation Protocol=OHEM, Sliding window2024.03 | 74.9 | |
| CCVCEvaluation Protocol=OHEM, Sliding window2024.03 | 74.9 | |
| U2PLVenue=CVPR'22, Encoder=RN-101, #Params=59.5M2024.10 | 74.9 | |
| AELEvaluation Protocol=OHEM, Sliding window2024.03 | 74.45 | |
| iMASVenue=CVPR'23, Encoder=RN-101, #Params=59.5M2024.10 | 74.3 | |
| PCRBackbone=ResNet-1012023.11 | 73.4 | |
| PCRVenue=NeurIPS'22, Encoder=RN-101, #Params=59.5M2024.10 | 73.4 | |
| DGCLEvaluation Protocol=OHEM, Sliding window2024.03 | 73.18 | |
| TQDMBB=ViT-B2025.06 | 72.8 | |
| SemiCVTEvaluation Protocol=OHEM, Sliding window2024.03 | 72.19 | |
| USRNBackbone=ResNet-50, Crop Size=712, Partition Protocol=1/162023.03 | 71.2 | |
| CPSBackbone=ResNet-50, Crop Size=712, Partition Protocol=1/16, reproduced by=U2PL2023.03 | 69.8 | |
| CPSBackbone=ResNet-1012023.11 | 69.8 | |
| CPSEvaluation Protocol=OHEM, Sliding window2024.03 | 69.78 | |
| GTA-SegVenue=NeurIPS'22, Encoder=RN-101, #Params=59.5M2024.10 | 69.4 | |
| GTA-SegEvaluation Protocol=OHEM, Sliding window2024.03 | 69.38 | |
| U2PLBackbone=ResNet-50, Crop Size=712, Partition Protocol=1/162023.03 | 69 | |
| ST++Evaluation Protocol=OHEM, Sliding window2024.03 | 67.64 | |
| CCTBackbone=ResNet-50, Crop Size=712, Partition Protocol=1/162023.03 | 66.4 | |
| SupBaseline-CNNBackbone=ResNet-101, Decoder=DeepLabV3+, Evaluation Protocol=OHEM, Sliding window2024.03 | 66.3 | |
| Labeled OnlyEncoder=RN-101, #Params=59.5M2024.10 | 66.3 | |
| GCTBackbone=ResNet-50, Crop Size=712, Partition Protocol=1/162023.03 | 65.8 | |
| SupBaseline-ViTBackbone=SegFormer-B5, Evaluation Protocol=OHEM, Sliding window2024.03 | 65.69 | |
| Supervised BaselineBackbone=ResNet-50, Crop Size=712, Partition Protocol=1/162023.03 | 63.3 |