List your company
ResearchFormerly WizWand

Generalized Few-Shot Semantic Segmentation on PASCAL-5^i (test)

76.42Base mIoUOurs (w/ trans.)
5.59623.98342.3760.757Nov 25, 2022Feb 17, 2023May 13, 2023Aug 6, 2023Oct 30, 2023Jan 23, 2024Apr 17, 2024
Updated 8mo ago

Evaluation results

MethodLinks
Ours (w/ trans.)Learning=Transductive, Backbone=ResNet-502024.0476.4256.1266.27
Ours (w/o trans.)Learning=Inductive, Backbone=ResNet-502024.0474.8650.3462.6
POPLearning=Transductive, Backbone=ResNet-502024.0474.7855.8765.33
BAMPractical setting=false, Multi-class design=true, Shots=1, Backbone=ResNet-502022.1171.627.4949.55
BAMPractical setting=false, Multi-class design=true, Shots=5, Backbone=ResNet-502022.1171.628.9650.28
BAMLearning=Inductive, Backbone=ResNet-502024.0471.628.9650.28
DIaM-UBPractical setting=false, Multi-class design=true, Shots=1, Backbone=ResNet-502022.1171.1352.6161.87
DIaM-UBPractical setting=false, Multi-class design=true, Shots=5, Backbone=ResNet-502022.1171.1266.1268.62
DIaMPractical setting=true, Multi-class design=true, Shots=1, Backbone=ResNet-502022.1170.8935.1153
DIaMPractical setting=true, Multi-class design=true, Shots=5, Backbone=ResNet-502022.1170.8555.3163.08
DIaMLearning=Transductive, Backbone=ResNet-502024.0470.8555.3163.08
MiBPractical setting=true, Multi-class design=false, Shots=5, Backbone=ResNet-502022.1168.628.9348.77
MiBLearning=Inductive, Backbone=ResNet-502024.0468.628.9348.77
CAPLPractical setting=false, Multi-class design=true, Shots=5, Backbone=ResNet-502022.1165.4324.4344.93
CAPLLearning=Inductive, Backbone=ResNet-502024.0465.4324.4344.93
CAPLPractical setting=false, Multi-class design=true, Shots=1, Backbone=ResNet-502022.1164.817.4641.13
DIaM (w/o trans.)Learning=Inductive, Backbone=ResNet-502024.0464.0534.5649.31
MiBPractical setting=true, Multi-class design=false, Shots=1, Backbone=ResNet-502022.1163.88.8636.33
POPLearning=Inductive, Backbone=ResNet-502024.0441.536.2638.8
RePRIPractical setting=true, Multi-class design=false, Shots=5, Backbone=ResNet-502022.1134.0620.9827.52
RePRILearning=Transductive, Backbone=ResNet-502024.0434.0620.9827.52
PANetPractical setting=true, Multi-class design=false, Shots=5, Backbone=ResNet-502022.1132.9515.2524.1
PANETLearning=Inductive, Backbone=ResNet-502024.0432.9515.2524.1
PANetPractical setting=true, Multi-class design=false, Shots=1, Backbone=ResNet-502022.1131.8811.2521.57
RePRIPractical setting=true, Multi-class design=false, Shots=1, Backbone=ResNet-502022.1120.7610.515.63
SCLPractical setting=true, Multi-class design=false, Shots=5, Backbone=ResNet-502022.119.111.835.47
SCLLearning=Inductive, Backbone=ResNet-502024.049.111.835.47
CANetPractical setting=true, Multi-class design=false, Shots=5, Backbone=ResNet-502022.119.051.525.29
CANeTLearning=Inductive, Backbone=ResNet-502024.049.051.525.29
SCLPractical setting=true, Multi-class design=false, Shots=1, Backbone=ResNet-502022.118.882.445.66
PFENetPractical setting=true, Multi-class design=false, Shots=5, Backbone=ResNet-502022.118.831.895.36
PFENETLearning=Inductive, Backbone=ResNet-502024.048.831.895.36
CANetPractical setting=true, Multi-class design=false, Shots=1, Backbone=ResNet-502022.118.732.425.58
PFENetPractical setting=true, Multi-class design=false, Shots=1, Backbone=ResNet-502022.118.322.675.5