Semantic Segmentation on COCO-Stuff 164K (test)
47.3mIoUCAA
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| CAABackbone=EfficientNet-B5, Multi-scale=true2021.01 | 47.3 | — | — | — | — | — | — | — | |
| Segformer-B5Multi-scale=true2021.01 | 46.7 | — | — | — | — | — | — | — | |
| SCASegBackbone=MiT-B2, Params. (M)=35.2, FPS (img/s)=27.00, GFLOPs=39.62024.11 | 45.85 | — | — | — | — | — | — | — | |
| SCASegBackbone=MiT-B2, Params. (M)=35.2, FPS (img/s)=27.00, GFLOPs=39.62024.11 | 45.85 | — | — | — | — | — | — | — | |
| SETRMulti-scale=true2021.01 | 45.8 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025 WACV, Backbone=MiT-B2, Params. (M)=35.8, FPS (img/s)=26.34, GFLOPs=40.02024.11 | 45.47 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025, Backbone=MiT-B2, Params. (M)=35.8, FPS (img/s)=26.34, GFLOPs=40.02024.11 | 45.47 | — | — | — | — | — | — | — | |
| VWFormerYear=2024 ICLR, Backbone=MiT-B2, Params. (M)=27.4, FPS (img/s)=27.68, GFLOPs=46.62024.11 | 45.18 | — | — | — | — | — | — | — | |
| VWFormerYear=2024, Backbone=MiT-B2, Params. (M)=27.4, FPS (img/s)=27.68, GFLOPs=46.62024.11 | 45.18 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023 AAAI, Backbone=MiT-B2, Params. (M)=29.1, FPS (img/s)=27.56, GFLOPs=42.72024.11 | 44.79 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023, Backbone=MiT-B2, Params. (M)=29.1, FPS (img/s)=27.56, GFLOPs=42.72024.11 | 44.79 | — | — | — | — | — | — | — | |
| MetaSegYear=2024 WACV, Backbone=MiT-B2, Params. (M)=27.8, FPS (img/s)=28.47, GFLOPs=25.22024.11 | 44.5 | — | — | — | — | — | — | — | |
| MetaSegYear=2024, Backbone=MiT-B2, Params. (M)=27.8, FPS (img/s)=28.47, GFLOPs=25.22024.11 | 44.5 | — | — | — | — | — | — | — | |
| SCASegBackbone=MSCAN-S, Params. (M)=23.7, FPS (img/s)=27.71, GFLOPs=20.32024.11 | 43.65 | — | — | — | — | — | — | — | |
| SCASegBackbone=MSCAN-S, Params. (M)=23.7, FPS (img/s)=27.71, GFLOPs=20.32024.11 | 43.65 | — | — | — | — | — | — | — | |
| SegFormerYear=2021 NeurIPS, Backbone=MiT-B2, Params. (M)=27.5, FPS (img/s)=29.72, GFLOPs=62.42024.11 | 43.42 | — | — | — | — | — | — | — | |
| SegFormerYear=2021, Backbone=MiT-B2, Params. (M)=27.5, FPS (img/s)=29.72, GFLOPs=62.42024.11 | 43.42 | — | — | — | — | — | — | — | |
| SCASegBackbone=MiT-B1, Params. (M)=23.4, FPS (img/s)=25.37, GFLOPs=17.42024.11 | 43.19 | — | — | — | — | — | — | — | |
| SCASegBackbone=MiT-B1, Params. (M)=23.4, FPS (img/s)=25.37, GFLOPs=17.42024.11 | 43.19 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025 WACV, Backbone=MSCAN-S, Params. (M)=24.3, FPS (img/s)=26.55, GFLOPs=20.82024.11 | 42.91 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025, Backbone=MSCAN-S, Params. (M)=24.3, FPS (img/s)=26.55, GFLOPs=20.82024.11 | 42.91 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025 WACV, Backbone=MiT-B1, Params. (M)=24.0, FPS (img/s)=22.06, GFLOPs=17.82024.11 | 42.71 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025, Backbone=MiT-B1, Params. (M)=24.0, FPS (img/s)=22.06, GFLOPs=17.82024.11 | 42.71 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023 AAAI, Backbone=MSCAN-S, Params. (M)=17.6, FPS (img/s)=24.16, GFLOPs=23.62024.11 | 42.61 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023, Backbone=MSCAN-S, Params. (M)=17.6, FPS (img/s)=24.16, GFLOPs=23.62024.11 | 42.61 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023 AAAI, Backbone=MiT-B1, Params. (M)=17.3, FPS (img/s)=22.29, GFLOPs=20.72024.11 | 42.42 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023, Backbone=MiT-B1, Params. (M)=17.3, FPS (img/s)=22.29, GFLOPs=20.72024.11 | 42.42 | — | — | — | — | — | — | — | |
| HRNetV2 + OCRBackbone=HRNetV2, Multi-scale=true2021.01 | 42.3 | — | — | — | — | — | — | — | |
| SCASegBackbone=SegMAN-T, Params. (M)=4.3, FPS (img/s)=43.02, GFLOPs=6.92024.11 | 42.22 | — | — | — | — | — | — | — | |
| SCASegBackbone=SegMAN-T, Params. (M)=4.3, FPS (img/s)=43.02, GFLOPs=6.92024.11 | 42.22 | — | — | — | — | — | — | — | |
| MetaSegYear=2024 WACV, Backbone=MSCAN-S, Params. (M)=16.3, FPS (img/s)=29.83, GFLOPs=15.32024.11 | 42.13 | — | — | — | — | — | — | — | |
| MetaSegYear=2024, Backbone=MSCAN-S, Params. (M)=16.3, FPS (img/s)=29.83, GFLOPs=15.32024.11 | 42.13 | — | — | — | — | — | — | — | |
| MetaSegYear=2024 WACV, Backbone=MiT-B1, Params. (M)=16.0, FPS (img/s)=27.12, GFLOPs=12.42024.11 | 42.04 | — | — | — | — | — | — | — | |
| MetaSegYear=2024, Backbone=MiT-B1, Params. (M)=16.0, FPS (img/s)=27.12, GFLOPs=12.42024.11 | 42.04 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025 WACV, Backbone=SegMAN-T, Params. (M)=4.3, FPS (img/s)=41.93, GFLOPs=7.02024.11 | 41.98 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025, Backbone=SegMAN-T, Params. (M)=4.3, FPS (img/s)=41.93, GFLOPs=7.02024.11 | 41.98 | — | — | — | — | — | — | — | |
| VWFormerYear=2024 ICLR, Backbone=MSCAN-S, Params. (M)=15.5, FPS (img/s)=27.37, GFLOPs=22.52024.11 | 41.76 | — | — | — | — | — | — | — | |
| VWFormerYear=2024, Backbone=MSCAN-S, Params. (M)=15.5, FPS (img/s)=27.37, GFLOPs=22.52024.11 | 41.76 | — | — | — | — | — | — | — | |
| MetaSegYear=2024 WACV, Backbone=SegMAN-T, Params. (M)=2.9, FPS (img/s)=42.81, GFLOPs=4.52024.11 | 41.68 | — | — | — | — | — | — | — | |
| MetaSegYear=2024, Backbone=SegMAN-T, Params. (M)=2.9, FPS (img/s)=42.81, GFLOPs=4.52024.11 | 41.68 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023 AAAI, Backbone=SegMAN-T, Params. (M)=3.1, FPS (img/s)=41.66, GFLOPs=7.92024.11 | 41.54 | — | — | — | — | — | — | — | |
| VWFormerYear=2024 ICLR, Backbone=MiT-B1, Params. (M)=13.7, FPS (img/s)=23.87, GFLOPs=13.22024.11 | 41.54 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023, Backbone=SegMAN-T, Params. (M)=3.1, FPS (img/s)=41.66, GFLOPs=7.92024.11 | 41.54 | — | — | — | — | — | — | — | |
| VWFormerYear=2024, Backbone=MiT-B1, Params. (M)=13.7, FPS (img/s)=23.87, GFLOPs=13.22024.11 | 41.54 | — | — | — | — | — | — | — | |
| SegNeXtYear=2022 NeurIPS, Backbone=MSCAN-S, Params. (M)=13.9, FPS (img/s)=30.91, GFLOPs=15.92024.11 | 41.42 | — | — | — | — | — | — | — | |
| SegNeXtYear=2022, Backbone=MSCAN-S, Params. (M)=13.9, FPS (img/s)=30.91, GFLOPs=15.92024.11 | 41.42 | — | — | — | — | — | — | — | |
| VWFormerYear=2024 ICLR, Backbone=SegMAN-T, Params. (M)=2.8, FPS (img/s)=42.15, GFLOPs=6.32024.11 | 41.38 | — | — | — | — | — | — | — | |
| VWFormerYear=2024, Backbone=SegMAN-T, Params. (M)=2.8, FPS (img/s)=42.15, GFLOPs=6.32024.11 | 41.38 | — | — | — | — | — | — | — | |
| SegMANYear=2025 CVPR, Backbone=SegMAN-T, Params. (M)=6.4, FPS (img/s)=44.52, GFLOPs=6.22024.11 | 41.3 | — | — | — | — | — | — | — | |
| SegMANYear=2025, Backbone=SegMAN-T, Params. (M)=6.4, FPS (img/s)=44.52, GFLOPs=6.22024.11 | 41.3 | — | — | — | — | — | — | — | |
| SegFormerYear=2021 NeurIPS, Backbone=MiT-B1, Params. (M)=13.7, FPS (img/s)=31.44, GFLOPs=15.92024.11 | 40.97 | — | — | — | — | — | — | — | |
| SegFormerYear=2021, Backbone=MiT-B1, Params. (M)=13.7, FPS (img/s)=31.44, GFLOPs=15.92024.11 | 40.97 | — | — | — | — | — | — | — | |
| SCASegBackbone=MSCAN-T, Params. (M)=6.5, FPS (img/s)=29.04, GFLOPs=7.42024.11 | 40.89 | — | — | — | — | — | — | — | |
| SCASegBackbone=MSCAN-T, Params. (M)=6.5, FPS (img/s)=29.04, GFLOPs=7.42024.11 | 40.89 | — | — | — | — | — | — | — | |
| SCASegBackbone=MiT-B0, Params. (M)=6.0, FPS (img/s)=40.25, GFLOPs=5.92024.11 | 40.56 | — | — | — | — | — | — | — | |
| SCASegBackbone=MiT-B0, Params. (M)=6.0, FPS (img/s)=40.25, GFLOPs=5.92024.11 | 40.56 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025 WACV, Backbone=MiT-B0, Params. (M)=6.1, FPS (img/s)=38.94, GFLOPs=6.12024.11 | 40.24 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025, Backbone=MiT-B0, Params. (M)=6.1, FPS (img/s)=38.94, GFLOPs=6.12024.11 | 40.24 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025 WACV, Backbone=MSCAN-T, Params. (M)=6.7, FPS (img/s)=27.69, GFLOPs=7.62024.11 | 40.04 | — | — | — | — | — | — | — | |
| U-MixFormerYear=2025, Backbone=MSCAN-T, Params. (M)=6.7, FPS (img/s)=27.69, GFLOPs=7.62024.11 | 40.04 | — | — | — | — | — | — | — | |
| MetaSegYear=2024 WACV, Backbone=MSCAN-T, Params. (M)=4.7, FPS (img/s)=31.74, GFLOPs=5.52024.11 | 39.7 | — | — | — | — | — | — | — | |
| MetaSegYear=2024, Backbone=MSCAN-T, Params. (M)=4.7, FPS (img/s)=31.74, GFLOPs=5.52024.11 | 39.7 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023 AAAI, Backbone=MSCAN-T, Params. (M)=5.0, FPS (img/s)=25.98, GFLOPs=9.32024.11 | 39.39 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023, Backbone=MSCAN-T, Params. (M)=5.0, FPS (img/s)=25.98, GFLOPs=9.32024.11 | 39.39 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023 AAAI, Backbone=MiT-B0, Params. (M)=4.5, FPS (img/s)=34.80, GFLOPs=7.82024.11 | 39.03 | — | — | — | — | — | — | — | |
| FeedFormerYear=2023, Backbone=MiT-B0, Params. (M)=4.5, FPS (img/s)=34.80, GFLOPs=7.82024.11 | 39.03 | — | — | — | — | — | — | — | |
| VWFormerYear=2024 ICLR, Backbone=MSCAN-T, Params. (M)=5.8, FPS (img/s)=28.06, GFLOPs=13.62024.11 | 38.85 | — | — | — | — | — | — | — | |
| VWFormerYear=2024, Backbone=MSCAN-T, Params. (M)=5.8, FPS (img/s)=28.06, GFLOPs=13.62024.11 | 38.85 | — | — | — | — | — | — | — | |
| SegNeXtYear=2022 NeurIPS, Backbone=MSCAN-T, Params. (M)=4.3, FPS (img/s)=33.16, GFLOPs=6.62024.11 | 38.7 | — | — | — | — | — | — | — | |
| SegNeXtYear=2022, Backbone=MSCAN-T, Params. (M)=4.3, FPS (img/s)=33.16, GFLOPs=6.62024.11 | 38.7 | — | — | — | — | — | — | — | |
| DeepLabV3+Backbone=ResNet-50, Multi-scale=true2021.01 | 38.4 | — | — | — | — | — | — | — | |
| MetaSegYear=2024 WACV, Backbone=MiT-B0, Params. (M)=4.1, FPS (img/s)=42.64, GFLOPs=3.92024.11 | 38.25 | — | — | — | — | — | — | — | |
| MetaSegYear=2024, Backbone=MiT-B0, Params. (M)=4.1, FPS (img/s)=42.64, GFLOPs=3.92024.11 | 38.25 | — | — | — | — | — | — | — | |
| VWFormerYear=2024 ICLR, Backbone=MiT-B0, Params. (M)=3.7, FPS (img/s)=39.63, GFLOPs=5.82024.11 | 36.28 | — | — | — | — | — | — | — | |
| VWFormerYear=2024, Backbone=MiT-B0, Params. (M)=3.7, FPS (img/s)=39.63, GFLOPs=5.82024.11 | 36.28 | — | — | — | — | — | — | — | |
| SegFormerYear=2021 NeurIPS, Backbone=MiT-B0, Params. (M)=3.8, FPS (img/s)=43.65, GFLOPs=8.42024.11 | 35.63 | — | — | — | — | — | — | — | |
| SegFormerYear=2021, Backbone=MiT-B0, Params. (M)=3.8, FPS (img/s)=43.65, GFLOPs=8.42024.11 | 35.63 | — | — | — | — | — | — | — | |
| DeepLab-V3+Head=DeepLab-V3+, Backbone=MobileNet-V2, fps=38.42, #params.=15.35M2023.03 | — | 31.18 | 32.01 | — | — | — | — | — | |
| DeepLab-V3+Head=DeepLab-V3+, Backbone=ResNet-50, fps=24.09, #params.=43.69M2023.03 | — | 40.85 | 41.49 | — | — | — | — | — | |
| DeepLab-V3+Head=DeepLab-V3+, Backbone=ResNet-101, fps=16.39, #params.=62.68M2023.03 | — | 42.39 | 42.96 | — | — | — | — | — | |
| DeepLab-V3+ + CACHead=DeepLab-V3+, Backbone=MobileNet-V2, fps=36.25, #params.=16.13M2023.03 | — | 34.71 | 35.81 | — | — | — | — | — | |
| DeepLab-V3+ + CACHead=DeepLab-V3+, Backbone=ResNet-50, fps=23.54, #params.=44.48M2023.03 | — | 42.99 | 43.97 | — | — | — | — | — | |
| DeepLab-V3+ + CACHead=DeepLab-V3+, Backbone=ResNet-101, fps=16.03, #params.=63.47M2023.03 | — | 44.21 | 45.1 | — | — | — | — | — | |
| FCNHead=FCN, Backbone=MobileNet-V2, fps=51.1, #params.=9.82M2023.03 | — | 15.28 | 17.01 | — | — | — | — | — | |
| FCN + CACHead=FCN, Backbone=MobileNet-V2, fps=49.46, #params.=10.61M2023.03 | — | 25.37 | 27.17 | — | — | — | — | — | |
| FreeSegEvaluation Setting=Transductive setting2024.03 | — | — | 42.2 | 49.1 | 45.3 | — | — | — | |
| Last LayerBackbone=DINOv2-S, Encoder=Frozen2026.05 | — | — | — | — | — | 30.77 | — | — | |
| Last LayerBackbone=DINOv3-S/16, Encoder=Frozen2026.05 | — | — | — | — | — | 31.15 | — | — | |
| Last LayerBackbone=BEiT-B/16, Encoder=Frozen2026.05 | — | — | — | — | — | 16.71 | — | — | |
| LawinEncoder=MiT-B5, Crop Size=512x512, Iterations=80K2022.12 | — | 47.5 | — | — | — | — | — | — | |
| LOESBackbone=DINOv2-S, Encoder=Frozen, k=3, Selected Layers=[0, 8, 11]2026.05 | — | — | — | — | — | — | 31.19 | 0.42 | |
| LOESBackbone=DINOv3-S/16, Encoder=Frozen, k=3, Selected Layers=[5, 9, 11]2026.05 | — | — | — | — | — | — | 31.16 | 0.01 | |
| LOESBackbone=BEiT-B/16, Encoder=Frozen, k=3, Selected Layers=[0, 3, 11]2026.05 | — | — | — | — | — | — | 17.83 | 1.12 | |
| MaskCLIP+Evaluation Setting=Transductive setting2024.03 | — | — | 39.6 | 54.7 | 45 | — | — | — | |
| MVP-SEG+Evaluation Setting=Transductive setting2024.03 | — | — | 39.9 | 55.8 | 45.5 | — | — | — | |
| OCRNetHead=OCRNet, Backbone=HRNet-W18, fps=14.62, #params.=12.18M2023.03 | — | 31.58 | 32.34 | — | — | — | — | — | |
| OCRNetHead=OCRNet, Backbone=HRNet-W48, fps=13.78, #params.=70.53M2023.03 | — | 40.4 | 41.66 | — | — | — | — | — | |
| OCRNet + CACHead=OCRNet, Backbone=HRNet-W18, fps=14.37, #params.=12.97M2023.03 | — | 39.12 | 40.65 | — | — | — | — | — | |
| OCRNet + CACHead=OCRNet, Backbone=HRNet-W48, fps=13.33, #params.=71.32M2023.03 | — | 42.64 | 43.53 | — | — | — | — | — | |
| OTSegEvaluation Setting=Inductive setting2024.03 | — | — | 41.4 | 41.4 | 41.4 | — | — | — |