List your company
ResearchFormerly WizWand

Image Classification on CUB-200-2011 (test)

91.9Top-1 AccSR-GNN (Full Model)
85.97287.51189.0590.589Apr 29, 2015Nov 5, 2016May 16, 2018Nov 24, 2019Jun 2, 2021Dec 11, 2022Jun 20, 2024
Updated 4mo ago

Evaluation results

MethodLinks
SR-GNN (Full Model)Param (M)=30.9, GFLOPS (B)=9.8, Infer. time per img. (ms)=5.02022.0991.9—4
TransFGBackbone=ViT-B_162021.0791.7——
FFVTBackbone=ViT-B_162021.0791.6——
ViTBackbone=ViT-B_162021.0790.8——
CutMixFT Strategy=-, Backbone=ViT-B/16@3842024.0690.52——
CutMixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Standard2024.0690.52——
Stacked LSTMBackbone CNN=GoogleNet2019.1190.4——
StackedLSTMBackbone=GoogleNet2021.0790.4——
Diff-Mix + CutMixFT Strategy=TI+DB, Backbone=ViT-B/16@3842024.0690.35——
Diff-Mix + CutMixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Textual Inversion + DreamBooth2024.0690.35——
MixupBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Standard2024.0690.32——
Diff-MixFT Strategy=TI+DB, Backbone=ViT-B/16@3842024.0690.05——
Diff-MixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Textual Inversion + DreamBooth2024.0690.05——
API-NetBackbone=DenseNet1612021.0790——
SaSPA + CutMixFT Strategy=No fine-tuning, Backbone=ViT-B/16@3842024.0689.92——
SaSPA + CutMixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.0689.92——
Spatial RNNBackbone CNN=M-Net/D-Net2019.1189.7——
SaSPAFT Strategy=No fine-tuning, Backbone=ViT-B/16@3842024.0689.7——
SaSPABackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.0689.7——
PMGBackbone=ResNet502021.0789.6——
Real-guidanceBackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.0689.54——
ME-ASNInfer. time per img. (ms)=33.92022.0989.5——
Real-filteringBackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.0689.49——
DA-fusionBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Textual Inversion2024.0689.4——
BaselineBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Standard2024.0689.37——
PAIRS+SingleAnnotations=GT + KP2018.0189.2——
FDLBackbone=DenseNet1612021.0789.1——
PAIRS OnlyAnnotations=GT + KP2018.0188.7——
iSQRT-COVBackbone CNN=ResNet-1012019.1188.7——
ELOPEBackbone CNN=ResNet-1012019.1188.5——
S3NBackbone=ResNet502021.0788.5——
OursBase Model=ResNet502021.0188.45——
CMNBackbone=ResNet-502020.1288.2——
CIN1-Stage=true, Backbone=ResNet-1012020.0388.1——
ACNetBackbone=ResNet-502020.1288.1——
ACNetBase Model=ResNet502021.0188.1——
CINBackbone=ResNet1012021.0788.1——
DBTNetBackbone=ResNet1012021.0788.1——
ACNetBackbone=ResNet502021.0788.1——
LIOBackbone=ResNet-502020.1288——
TASNBackbone=1xResnet-50, Resolution=High-resolution2019.0387.9——
TASNBackbone=ResNet-502020.1287.9——
TASNBase Model=ResNet502021.0187.9——
TASNParam (M)=37.3, GFLOPS (B)=21.9, Infer. time per img. (ms)=7.52022.0987.9——
DCLBackbone CNN=ResNet-502019.1187.8——
DCLBackbone=ResNet-502020.1287.8——
DCLBase Model=ResNet502021.0187.8——
DCLBackbone=ResNet502021.0787.8——
Cross-XBackbone=ResNet-502020.1287.7——
Cross-XBackbone=ResNet502021.0787.7——
ST-PrototPNetBackbone=ResNet50, Pre-training=iNaturalist2024.0487.63——
Diff-Mix + CutMixFT Strategy=TI+DB, Backbone=ResNet50@4482024.0687.56——
Diff-Mix + CutMixBackbone=ResNet50, Resolution=448x448, FT Strategy=Textual Inversion + DreamBooth2024.0687.56——
NTSBase Model=ResNet502021.0187.52——
Lam et al.Annotations=GT+KP2018.0187.5——
NTS-NetBackbone CNN=ResNet-502019.1187.5——
NTS1-Stage=false2020.0387.5——
CIN1-Stage=true, Backbone=ResNet-502020.0387.5——
NTS-NetBackbone=ResNet-502020.1287.5——
NTS-NetBackbone=ResNet502021.0787.5——
DFL-CNNBackbone CNN=ResNet-502019.1187.4——
ELOPEBackbone CNN=ResNet-502019.1187.4——
DFL-CNN1-Stage=false2020.0387.4——
DFL-CNNBackbone=ResNet-502020.1287.4——
DFLBase Model=ResNet502021.0187.4——
DFL-CNNBackbone=ResNet502021.0787.4——
OursBase Model=VGG192021.0187.34——
NTSNBackbone=3xResnet-50, Resolution=High-resolution2019.0387.3——
iSQRT-COV(8k)1-Stage=true2020.0387.3——
RGInfer. time per img. (ms)=23.82022.0987.3——
CutMixFT Strategy=-, Backbone=ResNet50@4482024.0687.23——
CutMixBackbone=ResNet50, Resolution=448x448, FT Strategy=Standard2024.0687.23——
ISEBackbone CNN=ResNet-502019.1187.2——
Diff-MixFT Strategy=TI+DB, Backbone=ResNet50@4482024.0687.16——
Diff-MixBackbone=ResNet50, Resolution=448x448, FT Strategy=Textual Inversion + DreamBooth2024.0687.16——
SaSPA + CutMixFT Strategy=No fine-tuning, Backbone=ResNet50@4482024.0687.14——
SaSPA + CutMixBackbone=ResNet50, Resolution=448x448, FT Strategy=No fine-tuning2024.0687.14——
TASNBackbone=3xVGG-19, Resolution=High-resolution2019.0387.1——
HBP1-Stage=true2020.0387.1——
CGNL1-Stage=true2020.0387——
CBAMBase Model=ResNet502021.0186.99——
SaSPAFT Strategy=No fine-tuning, Backbone=ResNet50@4482024.0686.92——
SaSPABackbone=ResNet50, Resolution=448x448, FT Strategy=No fine-tuning2024.0686.92——
SENetBase Model=ResNet502021.0186.78——
Real-guidanceBackbone=ResNet50, Resolution=448x448, FT Strategy=No fine-tuning2024.0686.71——
MixupBackbone=ResNet50, Resolution=448x448, FT Strategy=Standard2024.0686.68——
BaselineBackbone=ResNet50, Resolution=448x448, FT Strategy=Standard2024.0686.64——
MaxEntBackbone=DenseNet1612021.0786.6——
MAMCBackbone=1xResnet-50, Resolution=High-resolution2019.0386.5——
MA-CNNBackbone CNN=VGG-192019.1186.5——
OSME-MAMCBackbone CNN=ResNet-1012019.1186.5——
MAMC1-Stage=true2020.0386.5——
MA-CNN1-Stage=false2020.0386.5——
MA-CNNBackbone=VGG-192020.1286.5——
GMMBackbone CNN=VGG-192019.1186.3——
DA-fusionBackbone=ResNet50, Resolution=448x448, FT Strategy=Textual Inversion2024.0686.3——
MCPNetBackbone=ResNet50, Pre-training=iNaturalist2024.0486.28——
BoostCNNfeatures=B-CNN (VGG-D)2015.0486.2——
Cui et al.Annotations=GT2018.0186.2——
A3MAttribute=with annotation, Parts=without annotation2019.0186.2——