List your company

Semantic Segmentation on Pascal VOC blended 2012 (train)

80.9mIoUUniMatch
60.41265.73171.0576.369Aug 21, 2022Sep 22, 2022Oct 24, 2022Nov 25, 2022Dec 27, 2022Jan 28, 2023Mar 2, 2023

Evaluation results

MethodLinks
UniMatchBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/16, Data Selection=Prioritize high-quality set2022.0880.9———
UniMatchBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/42022.0879.2———
CCVCBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=512, epochs=402023.0379———
U2PLBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=512, reproduced_by_this_paper=true2023.0378.7———
PS-MTBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0378.7———
UniMatchBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/82022.0878.4———
CCVCBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=512, epochs=402023.0378.4———
PS-MTBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0378.2———
UniMatchBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/162022.0878.1———
UniMatchBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/16, Data Selection=Prioritize high-quality set2022.0878.1———
CPSBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0377.7———
U2PLBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=512, reproduced_by_this_paper=true2023.0377.6———
UniMatchBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/42022.0877.2———
U2PLBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/16, Data Selection=Prioritize high-quality set2022.0877.2———
CCVCBackbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=512, epochs=402023.0377.2———
UniMatchBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/82022.0877———
UniMatchBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/82022.0876.9———
UniMatchBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/42022.0876.8———
ST++Backbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/42022.0876.6———
ELNBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0376.6———
ST++Backbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0376.6———
UniMatchBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/162022.0876.5———
CPSBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0376.4———
PS-MTBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0376.4———
CCVCBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=512, epochs=402023.0376.4———
CACBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/42022.0876.3———
ST++Backbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/82022.0876.3———
CACBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0376.3———
ST++Backbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0376.3———
U2PLBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=512, reproduced_by_this_paper=true2023.0376.2———
UniMatchBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/42022.0876.1———
CCVCBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=512, epochs=402023.0376.1———
UniMatchBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/82022.0875.8———
UniMatchBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/162022.0875.8———
PS-MTBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0375.7———
PS-MTBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/162022.0875.5———
PS-MTBackbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0375.5———
ST++Backbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/42022.0875.4———
ST++Backbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0375.4———
ELNBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0375.1———
U2PLBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=512, reproduced_by_this_paper=true2023.0375.1———
CPSBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0374.9———
U2PLBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/16, Data Selection=Prioritize high-quality set2022.0874.7———
CACBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/82022.0874.6———
CACBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0374.6———
ELNBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0374.6———
UniMatchBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/162022.0874.5———
ST++Backbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/162022.0874.5———
CPSBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/162022.0874.5———
CPSBackbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0374.5———
ST++Backbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0374.5———
CCVCBackbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=512, epochs=402023.0374.5———
ST++Backbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/82022.0874.4———
U2PLBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/162022.0874.4———
ST++Backbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0374.4———
U2PLBackbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=512, reproduced_by_this_paper=true2023.0374.4———
CutMix-SegBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0374.3———
Supervised BaselineBackbone=ResNet-101, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0374.2———
CACBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/42022.0874———
CACBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0374———
CPSBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0373.7———
ELNBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0373.2———
SupBaselineBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/42022.0872.8———
PS-MTBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/162022.0872.8———
PS-MTBackbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0372.8———
CutMix-SegBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0372.7———
ST++Backbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/162022.0872.6———
CutMix-SegBackbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0372.6———
ST++Backbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0372.6———
CutMix-SegBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0372.5———
CACBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/82022.0872.4———
CACBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/162022.0872.4———
CACBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0372.4———
CACBackbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0372.4———
Supervised BaselineBackbone=ResNet-50, Partition Protocol=1/4, Number of Labelled Images=2646, Crop Size=5122023.0372.3———
CPSBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/162022.0872———
U2PLBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/162022.0872———
CPSBackbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0372———
U2PLBackbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=512, reproduced_by_this_paper=true2023.0372———
Supervised BaselineBackbone=ResNet-101, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0371.1———
SupBaselineBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/42022.0870.8———
CutMix-SegBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0370.7———
SupBaselineBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/16, Data Selection=Prioritize high-quality set2022.0870.6———
SupBaselineBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/82022.0870.4———
CACBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/162022.0870.1———
CACBackbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0370.1———
CutMix-SegBackbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0368.9———
Supervised BaselineBackbone=ResNet-50, Partition Protocol=1/8, Number of Labelled Images=1323, Crop Size=5122023.0368.2———
SupBaselineBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/16, Data Selection=Prioritize high-quality set2022.0867.7———
SupBaselineBackbone=ResNet-101, Training Resolution=513, Labeled Ratio=1/162022.0867.5———
Supervised BaselineBackbone=ResNet-101, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0367.5———
SupBaselineBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/82022.0867.3———
SupBaselineBackbone=ResNet-101, Training Resolution=321, Labeled Ratio=1/162022.0865.6———
SupBaselineBackbone=ResNet-50, Training Resolution=513, Labeled Ratio=1/162022.0862.4———
Supervised BaselineBackbone=ResNet-50, Partition Protocol=1/16, Number of Labelled Images=662, Crop Size=5122023.0362.4———
SupBaselineBackbone=ResNet-50, Training Resolution=321, Labeled Ratio=1/162022.0861.2———
AllSparkEncoder=MiT-B52024.05—78.38080.4
AugSegBackbone=ResNet-50, #Params=43.6M2023.10—74.6675.9977.16
CorrMatchEncoder=RN-1012024.05—77.677.878.3
CPSBackbone=ResNet-50, #Params=87.3M2023.10—71.9873.6774.9