List your company
ResearchFormerly WizWand

Scene Classification on AID TR=20%

97.49AccuracySelectiveMAE
49.899662.254874.6186.9652May 6, 2022Sep 11, 2022Jan 18, 2023May 27, 2023Oct 3, 2023Feb 9, 2024Jun 17, 2024
Updated 8mo ago

Evaluation results

MethodLinks
SelectiveMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=533k2024.0697.49
ViT-BPretrain=MAE, Backbone=ViT-B, Method=—2022.0897.47
SelectiveMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=533k, Pre-training=OpticalRS-13 (4M)2024.0697.25
ViTAE-BPretrain=MAE, Backbone=ViTAE-B, Method=—2022.0897.2
SelectiveMAEBackbone=ViT-B [2], Params (M)=86, Throughput/Minute=556k2024.0697.1
LSKNet-S#P=14.4M, FLOPS=54.4G2024.0397.05
RVSABackbone=VIT-B+RVSA [21], Params (M)=862024.0697.03
ViTAE-B + RVSAPretrain=MAE, Backbone=ViTAE-B + RVSA, Method=—2022.0897.03
RVSA#P=114.4M, FLOPS=301.3G2024.0397.01
ViTAE-B + RVSA♦Pretrain=MAE, Backbone=ViTAE-B + RVSA*, Method=—2022.0897.01
ViT-B + RVSAPretrain=MAE, Backbone=ViT-B + RVSA, Method=—2022.0896.92
RSP-VITAE#P=19.3M, FLOPS=119.1G2024.0396.91
VITAEv2-SPretrain=RSP, Backbone=VITAEv2-S, Method=—2022.0896.91
RingMoBackbone=Swin-B [102], Params (M)=882024.0696.9
SelectiveMAEBackbone=ViT-B [2], Params (M)=86, Throughput/Minute=556k, Pre-training=OpticalRS-13 (4M)2024.0696.9
RingMoPretrain=RingMo [32], Backbone=Swin-B, Method=—2022.0896.9
ViTAE-B + VSAPretrain=MAE, Backbone=ViTAE-B + VSA, Method=—2022.0896.9
ViT-B + RVSA♦Pretrain=MAE, Backbone=ViT-B + RVSA*, Method=—2022.0896.86
ViT-B + VSAPretrain=MAE, Backbone=ViT-B + VSA, Method=—2022.0896.85
RSP-Swin#P=27.5M, FLOPS=37.7G2024.0396.83
RSP-R50#P=25.6M, FLOPS=86.3G2024.0396.81
LSKNet-T#P=4.3M, FLOPS=19.2G2024.0396.8
CSPTPretrain=CSPT [80], Backbone=ViT-B, Method=—2022.0896.75
OREOLEBackbone=VIT-G [43], Params (M)=9142024.0696.71
VITAEv2-SPretrain=IMP, Backbone=VITAEv2-S, Method=—2022.0896.61
MAEBackbone=ViT-B [2], Params (M)=86, Throughput/Minute=264k, Pre-training=OpticalRS-13 (4M)2024.0696.58
Swin-TPretrain=IMP, Backbone=Swin-T, Method=—2022.0896.55
RingMoPretrain=RingMo [32], Backbone=ViT-B, Method=—2022.0896.54
FENet#P=23.9M, FLOPS=92.0G2024.0396.45
MGML-FENetPretrain=IMP, Backbone=DenseNet-121, Method=MGML-FENet [77]2022.0896.45
ScaleMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=206k2024.0696.44
CSPTPretrain=CSPT [80], Backbone=ViT-L, Method=—2022.0896.3
UPetu#P=87.7M, FLOPS=>322.2G2024.0396.29
CTNet2024.0396.25
F2BRBM#P=25.6M, FLOPS=86.3G2024.0396.05
F2BRBMPretrain=IMP, Backbone=ResNet-50, Method=F2BRBM [5]2022.0896.05
MBENet#P=23.9M, FLOPS=108.5G2024.0396
AGOSType=CNN, Backbone=DenseNet-1212022.0595.81
CADNetType=CNN, Year=20202022.0595.73
MBLANet2024.0395.6
MBLANetType=CNN, Year=20212022.0595.6
FSCNet#P=28.8M, FLOPS=166.1G2024.0395.56
AGOSType=CNN, Backbone=ResNet-1012022.0595.54
KFBNet2024.0395.5
GFMBackbone=Swin-B [102], Params (M)=882024.0695.47
GRMANet#P=54.1M, FLOPS=171.4G2024.0395.43
ConvNext#P=28.0M, FLOPS=93.7G2024.0395.43
GRMANetPretrain=IMP, Backbone=ResNet-50, Method=GRMANet [73]2022.0895.43
ASPPretrain=ASP [76], Backbone=ResNet-101, Method=—2022.0895.4
TOVBackbone=ResNet-50 [101], Params (M)=262024.0695.16
SatMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=205k2024.0695.02
AGOSType=CNN, Backbone=ResNet-502022.0594.99
SatLasBackbone=Swin-B [102], Params (M)=88, Throughput/Minute=243k2024.0694.96
IDCCP#P=25.6M, FLOPS=86.3G2024.0394.8
LSENet#P=25.9M, FLOPS=>86.3G2024.0394.41
LSENetPretrain=IMP, Backbone=VGG-16, Method=LSENet [42]2022.0894.41
EAM#P=>42.3M, FLOPS=>164.32024.0394.26
EAMPretrain=IMP, Backbone=ResNet-101, Method=EAM [74]2022.0894.26
LiGNetType=CNN, Year=20212022.0594.17
LSENetType=CNN, Year=20212022.0594.07
DSENetType=CNN, Year=20212022.0594.02
DMSMILType=CNN, Year=20212022.0593.98
MF2NetType=CNN, Year=20202022.0593.82
ViT-B#P=86.0M, FLOPS=118.9G2024.0393.81
ViT-BPretrain=IMP, Backbone=ViT-B, Method=—2022.0893.81
GASSLBackbone=ResNet-50 [101], Params (M)=262024.0693.55
MSANet#P=>42.3M, FLOPS=>164.32024.0393.53
MSANetPretrain=IMP, Backbone=ResNet-101, Method=MSANet [75]2022.0893.53
SeCoBackbone=ResNet-50 [101], Params (M)=262024.0693.47
MSDFFType=CNN, Year=20202022.0593.47
MG-CAP#P=>42.3M, FLOPS=>164.3G2024.0393.34
MG-CAPType=CNN, Year=20202022.0593.34
SCCov#P=13.0M2024.0393.12
GBNetType=CNN, Year=20202022.0592.2
MS2APType=CNN, Year=20212022.0592.19
MSCPType=CNN, Year=20182022.0591.52
SSL4EOBackbone=ViT-S [2], Params (M)=222024.0691.06
RBFFPretrain=IMP, Backbone=MobileNet-V2 [78], Method=RBFF [79]2022.0891.02
CACOBackbone=ResNet-50 [101], Params (M)=262024.0690.88
ARCNetType=RNN, Year=20182022.0588.75
APNetType=CNN, Year=20192022.0588.56
MIDCNetType=CNN, Year=20202022.0588.26
RANetType=CNN, Year=20202022.0588.12
SPPNetType=CNN, Year=2017, Note=Conducted by authors2022.0587.44
TEXNetType=CNN, Year=20172022.0587.32
AlexNetType=CNN, Year=20172022.0586.86
VGGNet-16Type=CNN, Year=20172022.0586.59
GoogLeNetType=CNN, Year=20172022.0583.44
AGANType=GAN, Year=20192022.0578.95
MARTAType=GAN, Year=20172022.0575.39
BOVW(SIFT)Type=Hand-crafted, Year=20172022.0562.49
PLSA(SIFT)Type=Hand-crafted, Year=20172022.0556.24
LDA(SIFT)Type=Hand-crafted, Year=20172022.0551.73