List your company
ResearchFormerly WizWand

Semantic Segmentation on Cityscapes fine annotations 1/2 (1488 labels)

84.8mIoUUniMatchV2-B + PrevMatch
76.4878.6480.882.96Nov 27, 2023Dec 28, 2023Jan 28, 2024Feb 28, 2024Mar 30, 2024Apr 30, 2024May 31, 2024
Updated 8mo ago

Evaluation results

MethodLinks
UniMatchV2-B + PrevMatchBackbone=DINOv2-B, Resolution=756x7562024.0584.8
UniMatchV2-BBackbone=DINOv2-B, Resolution=756x7562024.0583.9
UniMatchV2-S + PrevMatchBackbone=DINOv2-S2024.0583.2
UniMatchV2-SBackbone=DINOv2-S2024.0582.6
AllSparkBackbone=Transformer-based2024.0581.4
LogicDiagBackbone=ResNet-1012023.1181
AugSeg + PrevMatchBackbone=ResNet-1012024.0580.8
SemiVLBackbone=ViT-B/162023.1180.6
SemiVLBackbone=Transformer-based2024.0580.6
ESLBackbone=ResNet-1012023.1180.5
Dual TeacherBackbone=ResNet-1012024.0580.5
CorrMatchBackbone=ResNet-1012024.0580.4
AugSegBackbone=ResNet-1012024.0580.4
AELBackbone=ResNet-1012023.1180.3
UniMatch + PrevMatchBackbone=ResNet-1012024.0580.1
UniMatchBackbone=ViT-B/16, Note=re-produced2023.1179.6
UniMatchBackbone=ResNet-1012023.1179.5
UniMatchBackbone=ResNet-1012024.0579.5
U2PLBackbone=ResNet-1012023.1179.1
PCRBackbone=ResNet-1012023.1179.1
PS-MTBackbone=ResNet-1012024.0579.1
CPSBackbone=ResNet-1012023.1176.8