List your company
ResearchFormerly WizWand

Semantic Segmentation on NYU Depth V2

68.1mIoUInternImage-H
42.6249.23555.8562.465Nov 10, 2022Jun 14, 2023Jan 17, 2024Aug 20, 2024Mar 25, 2025Oct 27, 2025Jun 1, 2026
Updated 5mo ago

Evaluation results

MethodLinks
InternImage-HBackbone=InternImage-H2022.1168.1
OraclesBackbone=SED [71]2025.0360.6
SigmaBackbone=VMamba-S, Params=69.8, Input Size=480 x 6402024.0457
CMNeXtBackbone=MiT-B4, Params=119.6, Input Size=480 x 6402024.0456.9
CMX-B5Backbone=CMX-B52022.1156.9
MultiMAEBackbone=ViT-B, Params=95.2, Input Size=640 x 6402024.0456
MultiMAEBackbone=ViT-B, Params=95.2M2026.0556
SMMCLBackbone=SegNeXt-B, Input Size=480 x 6402024.0455.8
DFormerBackbone=DFormer-Base, Params=29.5M2026.0555.6
T-PrompterBackbone=T-Prompter-L, Params=401†, Input Size=480 x 6402024.0455.3
AsymFormerBackbone=MiT-B0+ConvNeXt-Tiny, Params=33.0M2026.0555.3
RelFlexformer (w/ DFormer)Backbone=DFormer-Base, Params=29.5M2026.0555.3
CMXBackbone=MiT-B2, Params=66.6M2026.0554.4
Performer (w/ DFormer)Backbone=DFormer-Base, Params=29.5M2026.0554.4
TokenFusionBackbone=MiT-B3, Params=45.9, Input Size=480 x 6402024.0454.2
TokenFusionBackbone=MiT-B3, Params=45.9M2026.0554.2
OmnivoreBackbone=Swin-B, Params=95.7, Input Size=480 x 6402024.0454
OmnivoreBackbone=Swin-Base, Params=95.7M2026.0554
SigmaBackbone=VMamba-T, Params=48.3, Input Size=480 x 6402024.0453.9
PGDENetBackbone=ResNet-34, Params=100.7, Input Size=480 x 6402024.0453.7
PGDENetBackbone=ResNet-34, Params=100.7M2026.0553.7
FRNetBackbone=ResNet-34, Params=85.5, Input Size=480 x 6402024.0453.6
InvPTBackbone=ViT-L, Params=423†, Input Size=480 x 6402024.0453.6
FRNetBackbone=ResNet-34, Params=85.5M2026.0553.6
DFormerBackbone=DFormer-Small, Params=18.7M2026.0553.6
PDCNetBackbone=ResNet-101, Input Size=480 x 4802024.0453.5
TokenFusionBackbone=MiT-B2, Params=26.0, Input Size=480 x 6402024.0453.3
TokenFusionBackbone=MiT-B2, Params=26.0M2026.0553.3
OmnivoreBackbone=Swin-Small, Params=51.3M2026.0552.7
CENBackbone=ResNet-152, Params=133.9, Input Size=480 x 6402024.0452.5
CENBackbone=ResNet-152, Params=133.9M2026.0552.5
SA-GateBackbone=ResNet-101, Params=110.9, Input Size=480 x 6402024.0452.4
SA-GateBackbone=ResNet-101, Params=110.9M2026.0552.4
CL-DMDFModel Category=Graph-Based, MAdds=43.42026.0652.3
SemLABackbone=SED [71]2025.0351.82
DFormerBackbone=DFormer-Tiny, Params=6.0M2026.0551.8
CENBackbone=ResNet-101, Params=118.2, Input Size=480 x 6402024.0451.7
CENBackbone=ResNet-101, Params=118.2M2026.0551.7
ShapeConvBackbone=ResNext-101, Params=86.8, Input Size=480 x 6402024.0451.3
ShapeConvBackbone=ResNet-101, Params=86.8M2026.0551.3
SGNetBackbone=ResNet-101, Params=64.7, Input Size=480 x 6402024.0451.1
SGNetBackbone=ResNet-101, Params=64.7M2026.0551.1
CENModel Category=Encoder-Decoder, MAdds=618.32026.0651.1
EMSANetBackbone=ResNet-34, Params=46.9, Input Size=480 x 6402024.0451
EMSANetBackbone=ResNet-34, Params=46.9M2026.0551
ESANetModel Category=Attention-Based, MAdds=56.92026.0650.6
MLTCModel Category=Graph-Based, MAdds=147.62026.0650.4
ESANetBackbone=ResNet-34, Params=31.2, Input Size=480 x 6402024.0450.3
ESANetBackbone=ResNet-34, Params=31.2M2026.0550.3
OmnivoreBackbone=Swin-Tiny, Params=29.1M2026.0549.7
UniformBackbone=SED [71]2025.0348.42
ACNetBackbone=ResNet-50, Params=116.6, Input Size=480 x 6402024.0448.3
ACNetBackbone=ResNet-50, Params=116.6M2026.0548.3
ACNetModel Category=Encoder-Decoder, MAdds=126.22026.0648.3
Zero-shotBackbone=SED [71]2025.0347.2
LW-RefineNetModel Category=Attention-Based, MAdds=38.52026.0643.6