Semantic Segmentation on PASCAL Context (val)
79.03mIoUInvPT
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| InvPT2022.03 | 79.03 | — | — | — | — | — | — | — | — | — | — | — | |
| ASTMT2022.03 | 68 | — | — | — | — | — | — | — | — | — | — | — | |
| ATRC-BMTAS2022.03 | 67.67 | — | — | — | — | — | — | — | — | — | — | — | |
| SegformerPre-training=Ours2022.02 | 65 | — | — | — | — | — | — | — | — | — | — | — | |
| UPerNet+IDRNetBackbone=Swin-Large, Stride=32x2023.10 | 64.51 | — | — | — | — | — | — | — | — | — | — | — | |
| Mask2FormerBackbone=Swin-L, Supervision=M2024.03 | 64.3 | — | — | — | — | — | — | — | — | — | — | — | |
| UPerNet+MCIBI++Backbone=Swin-Large, Stride=32x2023.10 | 64.01 | — | — | — | — | — | — | — | — | — | — | — | |
| ATRC-ASPP2022.03 | 63.6 | — | — | — | — | — | — | — | — | — | — | — | |
| TSGBackbone=Swin-Large, Params=250M2022.05 | 63.3 | — | — | — | — | — | — | — | — | — | — | — | |
| ATRC2022.03 | 62.69 | — | — | — | — | — | — | — | — | — | — | — | |
| SenFormerBackbone=Swin-Large, Params=314M2022.05 | 62.4 | — | — | — | — | — | — | — | — | — | — | — | |
| Cat-Seg+CLIPSelfModel=ViT-L/142023.10 | 62.3 | — | — | — | 80.7 | — | — | — | — | — | — | — | |
| Cat-SegModel=ViT-L/142023.10 | 62 | — | — | — | 78.3 | — | — | — | — | — | — | — | |
| MTI-Net2022.03 | 61.7 | — | — | — | — | — | — | — | — | — | — | — | |
| Mask2Former+IDRNetBackbone=Swin-Large, Stride=32x2023.10 | 61.65 | — | — | — | — | — | — | — | — | — | — | — | |
| SegformerPre-training=ADE20K2022.02 | 60.8 | — | — | — | — | — | — | — | — | — | — | — | |
| CAA2022.02 | 60.5 | — | — | — | — | — | — | — | — | — | — | — | |
| DPT2022.02 | 60.5 | — | — | — | — | — | — | — | — | — | — | — | |
| DPT-HybridBackbone=ViT-Hybrid2021.03 | 60.46 | — | — | 84.83 | — | — | — | — | — | — | — | — | |
| Swin Transformer + UperNetBackbone=Swin-Large, Params=234M2022.05 | 60.3 | — | — | — | — | — | — | — | — | — | — | — | |
| SegNeXt-LBackbone=MSCAN-Large, Stride=32x2023.10 | 60.3 | — | — | — | — | — | — | — | — | — | — | — | |
| OCR2022.02 | 59.6 | — | — | — | — | — | — | — | — | — | — | — | |
| SegformerPre-training=ImageNet2022.02 | 59.4 | — | — | — | — | — | — | — | — | — | — | — | |
| SegmenterBackbone=ViT-Large, Stride=16x2023.10 | 59 | — | — | — | — | — | — | — | — | — | — | — | |
| Seg-L-Mask/16Backbone=ViT-L/16, Multi-scale testing=true2021.05 | 59 | — | — | — | — | — | — | — | — | — | — | — | |
| Deeplab V3Backbone=ResNeSt-2692021.03 | 58.92 | — | — | 83.06 | — | — | — | — | — | — | — | — | |
| RegProxybackbone=ViT-L/16, FLOPs=183G, #params.=305M2022.03 | 58.4 | — | 58.8 | — | — | — | — | — | — | — | — | — | |
| Deeplab V3Backbone=ResNeSt-2002021.03 | 58.37 | — | — | 82.5 | — | — | — | — | — | — | — | — | |
| Segmenterbackbone=ViT-L/16, FLOPs=210G, #params.=333M2022.03 | 58.1 | — | 59 | — | — | — | — | — | — | — | — | — | |
| SegmenterBackbone=ViT-L, Params=333M2022.05 | 58.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Cat-Seg+CLIPSelfModel=ViT-B/162023.10 | 58 | — | — | — | 75.2 | — | — | — | — | — | — | — | |
| SANModel=ViT-L/142023.10 | 57.7 | — | — | — | 77.6 | — | — | — | — | — | — | — | |
| Cat-SegModel=ViT-B/162023.10 | 57.5 | — | — | — | 74 | — | — | — | — | — | — | — | |
| CondNetbackbone=ResNest-1012021.09 | 57 | — | — | — | — | — | — | — | — | — | — | — | |
| UPerNet+MCIBI++Backbone=ResNet-101, Stride=8x2023.10 | 56.82 | — | — | — | — | — | — | — | — | — | — | — | |
| DiffCutLanguage Dependency (LD)=false, Auxiliary Image (AX)=false, Unsupervised Adaptation (UA)=false, Training Type=Training-free2024.06 | 56.5 | — | — | — | — | — | — | — | — | — | — | — | |
| Seg-L/16Backbone=ViT-L/16, Multi-scale testing=true2021.05 | 56.5 | — | — | — | — | — | — | — | — | — | — | — | |
| OCNetBackbone=HRNet-W482021.03 | 56.2 | — | — | — | — | — | — | — | — | — | — | — | |
| OCRNetBackbone=HRNetV2-W48, Stride=4x2023.10 | 56.2 | — | — | — | — | — | — | — | — | — | — | — | |
| OCRBackbone=HRNetV2-W48, Multi-scale testing=true2021.05 | 56.2 | — | — | — | — | — | — | — | — | — | — | — | |
| CondNetbackbone=ResNet-1012021.09 | 56 | — | — | — | — | — | — | — | — | — | — | — | |
| SETRBackbone=ViT-Large, Stride=16x2023.10 | 55.83 | — | — | — | — | — | — | — | — | — | — | — | |
| STLNetBackbone=ResNet-1012021.03 | 55.8 | — | — | — | — | — | — | — | — | — | — | — | |
| SETR-L MLABackbone=ViT-L/16, Multi-scale testing=true2021.05 | 55.8 | — | — | — | — | — | — | — | — | — | — | — | |
| CaC-NetBackbone=ResNet-1012021.03 | 55.4 | — | — | — | — | — | — | — | — | — | — | — | |
| DRANetBackbone=ResNet-101, Multi-scale testing=true2021.05 | 55.4 | — | — | — | — | — | — | — | — | — | — | — | |
| DNLBackbone=HRNetV2-W48, Multi-scale testing=true2021.05 | 55.3 | — | — | — | — | — | — | — | — | — | — | — | |
| ODISEPre-training Dataset=COCO, Pre-training Supervision=mask2026.03 | 55.3 | — | — | — | — | — | — | — | — | — | — | — | |
| HamNetBackbone=ResNet-101, Output Stride=8, Optimization=NMF2021.09 | 55.2 | — | — | — | — | — | — | — | — | — | — | — | |
| RegProxybackbone=ViT-B/16, FLOPs=52G, #params.=87M2022.03 | 55.2 | — | 55.4 | — | — | — | — | — | — | — | — | — | |
| Seg-B†-Mask/16Backbone=DeiT-B/16, Multi-scale testing=true2021.05 | 55 | — | — | — | — | — | — | — | — | — | — | — | |
| SETRbackbone=ViT-L/16, FLOPs=238G, #params.=318M2022.03 | 54.9 | — | 55.8 | — | — | — | — | — | — | — | — | — | |
| SETR-MLABackbone=ViT-L, Params=311M2022.05 | 54.9 | — | — | — | — | — | — | — | — | — | — | — | |
| RecoNetBackbone=ResNet-1012021.03 | 54.8 | — | — | — | — | — | — | — | — | — | — | — | |
| OCRNetbackbone=ResNet-101, reference=ECCV20202021.09 | 54.8 | — | — | — | — | — | — | — | — | — | — | — | |
| OCRBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 54.8 | — | — | — | — | — | — | — | — | — | — | — | |
| SegformerBackbone=MiT-B5, Params=85M2022.05 | 54.8 | — | — | — | — | — | — | — | — | — | — | — | |
| OCRNetBackbone=ResNet-101, Stride=8x2023.10 | 54.8 | — | — | — | — | — | — | — | — | — | — | — | |
| OCNetBackbone=ResNet-101, Stride=8x2023.10 | 54.72 | — | — | — | — | — | — | — | — | — | — | — | |
| APCNetbackbone=ResNet-101, reference=CVPR20192021.09 | 54.7 | — | — | — | — | — | — | — | — | — | — | — | |
| APCNetBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 54.7 | — | — | — | — | — | — | — | — | — | — | — | |
| APCNetBackbone=ResNet-101, Stride=8x2023.10 | 54.7 | — | — | — | — | — | — | — | — | — | — | — | |
| APCNetBackbone=ResNet-101, Multi-scale testing=true2021.05 | 54.7 | — | — | — | — | — | — | — | — | — | — | — | |
| DeepLabv3+Backbone=ResNet-101, Supervision=M2024.03 | 54.6 | — | — | — | — | — | — | — | — | — | — | — | |
| TSGBackbone=Swin-Tiny, Params=72M2022.05 | 54.5 | — | — | — | — | — | — | — | — | — | — | — | |
| DMNetBackbone=ResNet-1012021.03 | 54.4 | — | — | — | — | — | — | — | — | — | — | — | |
| DMNetBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 54.4 | — | — | — | — | — | — | — | — | — | — | — | |
| ACNetBackbone=ResNet-101, Stride=8x2023.10 | 54.1 | — | — | — | — | — | — | — | — | — | — | — | |
| ACNetBackbone=ResNet-101, Multi-scale testing=true2021.05 | 54.1 | — | — | — | — | — | — | — | — | — | — | — | |
| CFNetBackbone=ResNet-1012021.03 | 54 | — | — | — | — | — | — | — | — | — | — | — | |
| CFNetbackbone=ResNet-101, reference=CVPR20192021.09 | 54 | — | — | — | — | — | — | — | — | — | — | — | |
| CFNetBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 54 | — | — | — | — | — | — | — | — | — | — | — | |
| CFNetBackbone=ResNet-101, Multi-scale testing=true2021.05 | 54 | — | — | — | — | — | — | — | — | — | — | — | |
| CPNetBackbone=ResNet-1012021.03 | 53.9 | — | — | — | — | — | — | — | — | — | — | — | |
| CPNetbackbone=ResNet-101, reference=CVPR20202021.09 | 53.9 | — | — | — | — | — | — | — | — | — | — | — | |
| RGNetbackbone=ResNet-101, reference=ECCV20202021.09 | 53.9 | — | — | — | — | — | — | — | — | — | — | — | |
| CPNetBackbone=ResNet-101, Multi-scale testing=true2021.05 | 53.9 | — | — | — | — | — | — | — | — | — | — | — | |
| Seg-B†/16Backbone=DeiT-B/16, Multi-scale testing=true2021.05 | 53.9 | — | — | — | — | — | — | — | — | — | — | — | |
| SANModel=ViT-B/162023.10 | 53.8 | — | — | — | 73 | — | — | — | — | — | — | — | |
| MaskFormerBackbone=ResNet101, Params=60M2022.05 | 53.7 | — | — | — | — | — | — | — | — | — | — | — | |
| PAD-Net2022.03 | 53.6 | — | — | — | — | — | — | — | — | — | — | — | |
| DHRBackbone=Swin-L, Supervision=I2024.03 | 53.6 | — | — | — | — | — | — | — | — | — | — | — | |
| OTSeg+Source Dataset=COCO-156, Evaluation Setting=Transductive2024.03 | 53.4 | — | — | — | — | — | — | — | — | — | — | — | |
| SVCNetBackbone=ResNet-1012021.03 | 53.2 | — | — | — | — | — | — | — | — | — | — | — | |
| SVCNetbackbone=ResNet-101, reference=CVPR20192021.09 | 53.2 | — | — | — | — | — | — | — | — | — | — | — | |
| SenFormerBackbone=Swin-Tiny, Params=144M2022.05 | 53.2 | — | — | — | — | — | — | — | — | — | — | — | |
| EncNet+JPUBackbone=ResNet-1012019.03 | 53.1 | — | — | — | — | — | — | — | — | — | — | — | |
| EMANetBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 53.1 | — | — | — | — | — | — | — | — | — | — | — | |
| EMANetBackbone=ResNet-101, Stride=8x2023.10 | 53.1 | — | — | — | — | — | — | — | — | — | — | — | |
| SANetBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 53 | — | — | — | — | — | — | — | — | — | — | — | |
| OTSegSource Dataset=COCO-156, Evaluation Setting=Transductive2024.03 | 52.9 | — | — | — | — | — | — | — | — | — | — | — | |
| SpyGRBackbone=ResNet-1012021.03 | 52.8 | — | — | — | — | — | — | — | — | — | — | — | |
| ANLbackbone=ResNet-101, reference=ICCV20192021.09 | 52.8 | — | — | — | — | — | — | — | — | — | — | — | |
| SpyGRBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 52.8 | — | — | — | — | — | — | — | — | — | — | — | |
| HRNetPre-training=Ours2022.02 | 52.8 | — | — | — | — | — | — | — | — | — | — | — | |
| ANNBackbone=ResNet-101, Multi-scale testing=true2021.05 | 52.8 | — | — | — | — | — | — | — | — | — | — | — | |
| DANetBackbone=ResNet-1012021.03 | 52.6 | — | — | — | — | — | — | — | — | — | — | — | |
| DANetbackbone=ResNet-101, reference=CVPR20192021.09 | 52.6 | — | — | — | — | — | — | — | — | — | — | — | |
| DANetBackbone=ResNet-101, Output Stride=8, Attention-based=true2021.09 | 52.6 | — | — | — | — | — | — | — | — | — | — | — | |
| HRNetPre-training=Mseg2022.02 | 52.6 | — | — | — | — | — | — | — | — | — | — | — |