List your company
ResearchFormerly WizWand

Semantic Segmentation on PASCAL-Context 60 classes (test)

55.5mIoUCTNet
16.60426.70236.846.898Apr 4, 2015Apr 6, 2016Apr 9, 2017Apr 12, 2018Apr 15, 2019Apr 17, 2020Apr 20, 2021
Updated 8mo ago

Evaluation results

MethodLinks
CTNetBackbone=ResNet-1012021.0455.5——
OCRBackbone=ResNet-1012021.0454.8——
APCNetBackbone=ResNet-1012021.0454.7——
SPNetBackbone=ResNet-1012021.0454.5——
ACNetBackbone=ResNet-1012021.0454.1——
CFNetBackbone=ResNet-1012021.0454——
CPNBackbone=ResNet-1012021.0453.9——
SFNetBackbone=ResNet-101, Inference Scale=multi-scale, #GFLOPs=93.62020.0253.8——
BAFPNetBackbone=ResNet-101, Inference Scale=multi-scale2020.0253.6——
SVCNetBackbone=ResNet-1012021.0453.2——
EMANetBackbone=ResNet-101, Inference Scale=multi-scale, #GFLOPs=209.32020.0253.1——
FastFCNBackbone=ResNet-1012021.0453.1——
EMANetBackbone=ResNet-1012021.0453.1——
CTNetBackbone=ResNet-502021.0452.9——
ANNetBackbone=ResNet-101, Inference Scale=multi-scale, #GFLOPs=243.82020.0252.8——
DANetBackbone=ResNet-101, Inference Scale=multi-scale, #GFLOPs=257.12020.0252.6——
DANetBackbone=ResNet-1012021.0452.6——
EncNetBaseNet=Res1012018.0351.7——
EncNetBackbone=ResNet-101, Inference Scale=multi-scale2020.0251.7——
EncNetBackbone=ResNet-1012021.0451.7——
Ding et al.Backbone=ResNet-101, Inference Scale=multi-scale2020.0251.6——
SFNet (w/o FAM)Backbone=ResNet-101, Inference Scale=multi-scale, #GFLOPs=92.72020.0251.1——
SFNetBackbone=ResNet-50, Inference Scale=multi-scale, #GFLOPs=75.42020.0250.7——
MSCIBackbone=ResNet-1522021.0450.3——
DANetBackbone=ResNet-50, Inference Scale=multi-scale, #GFLOPs=186.42020.0250.1——
EncNetBackbone=ResNet-50, Inference Scale=multi-scale2020.0249.2——
SFNet (w/o FAM)Backbone=ResNet-50, Inference Scale=multi-scale, #GFLOPs=74.52020.0249——
PSPNetBackbone=ResNet-1012021.0447.8——
RefineNet-Res152Backbone=ResNet-152, Extra train data=none2016.1147.3——
RefineNetBaseNet=Res1522018.0347.3——
RefineNetBackbone=ResNet-1522021.0447.3——
RefineNet-Res101Backbone=ResNet-101, Extra train data=none2016.1147.1——
DeepLab-v2Backbone=ResNet-101, Extra train data=COCO (~100K)2016.1145.7——
DeepLab-v2BaseNet=Res101-COCO2018.0345.7——
VeryDeep2018.0344.5——
VeryDeep2021.0444.5——
Contextual Deep CRFs2015.0443.371.553.9
ContextExtra train data=none2016.1143.3——
Piecewise2018.0343.3——
HO-CRFExtra train data=none2016.1141.3——
HO_CRF2018.0341.3——
BoxSup2015.0440.5——
BoxSupExtra train data=none2016.1140.5——
BoxSup2018.0340.5——
ParseNet2018.0340.4——
CRF-RNN2018.0339.3——
FCN-8s2018.0337.8——
FCN2021.0437.8——
FCN-8s2015.0435.165.946.5
FCN-8sExtra train data=none2016.1135.1——
CFM2015.0434.4——
CFMExtra train data=none2016.1134.4——
O2P2015.0418.1——
O2PExtra train data=none2016.1118.1——