List your company
ResearchFormerly WizWand

No-Reference Image Quality Assessment on LIVE

0.987SROCCOurs-S+
0.864280.896140.9280.95986Apr 19, 2022Dec 12, 2022Aug 7, 2023Mar 31, 2024Nov 24, 2024Jul 19, 2025Mar 14, 2026
Updated 7mo ago

Evaluation results

MethodLinks
Ours-S+Backbone=Swin-Tiny, Joint training (FR & NR tasks)=true, Architecture Type=Transformer-based2023.100.9870.987
Ours-SBackbone=Swin-Tiny, Joint training (FR & NR tasks)=false, Architecture Type=Transformer-based2023.100.9860.986
SHAMISAType=SSL+LR2026.030.9860.987
MANIQA2022.040.9820.983
MANIQAArchitecture Type=Transformer-based2023.100.9820.983
MAMIQA2026.020.9810.981
Ours-RBackbone=ResNet50, Joint training (FR & NR tasks)=false2023.100.9740.976
Su et al.Type=Supervised2026.030.9730.974
Re-IQAType=SSL+LR2026.030.970.971
TReS2022.040.9690.968
TRESArchitecture Type=CNN/Hand-crafted2023.100.9690.968
DBCNN2026.020.9690.971
TreS2026.020.9690.968
TReSType=Supervised2026.030.9690.968
DBCNN2022.040.9680.971
DBCNNArchitecture Type=CNN/Hand-crafted2023.100.9680.971
DB-CNNType=Supervised2026.030.9680.971
ARNIQAType=SSL+LR2026.030.9660.97
MS-SCANet2026.020.9640.968
HyperIQA2022.040.9620.966
HyperIQAArchitecture Type=CNN/Hand-crafted2023.100.9620.966
HyperIQA2026.020.9620.966
HyperIQAType=Supervised2026.030.9620.966
WaDIQaM2022.040.960.955
MetaIQA2022.040.960.959
WaDIQaMArchitecture Type=CNN/Hand-crafted2023.100.960.955
MetaIQAArchitecture Type=CNN/Hand-crafted2023.100.960.959
CONTRIQUEType=SSL+LR2026.030.960.961
P2P-BM2022.040.9590.958
P2P-BMArchitecture Type=CNN/Hand-crafted2023.100.9590.958
BIECON2022.040.9580.961
BIECONArchitecture Type=CNN/Hand-crafted2023.100.9580.961
MEON2022.040.9510.955
MEONArchitecture Type=CNN/Hand-crafted2023.100.9510.955
MEON2026.020.9510.955
TIQA2022.040.9490.965
TIQAArchitecture Type=Transformer-based2023.100.9490.965
TRIQ2026.020.9490.965
CORNIAType=Codebook2026.030.9470.95
HOSAType=Codebook2026.030.9460.95
SHAMISATraining Dataset=CSIQ2026.030.939—
BRISQUEType=Handcrafted2026.030.9390.935
Su et al.Training=CSIQ2023.100.93—
Su et al.Training Dataset=CSIQ2026.030.93—
BRISQUE2022.040.9290.944
BRISQUEArchitecture Type=CNN/Hand-crafted2023.100.9290.944
BRISQUE2026.020.9290.94
HyperIQATraining=CSIQ2023.100.926—
HyperIQATraining Dataset=CSIQ2026.030.926—
ARNIQATraining=CSIQ2023.100.921—
ARNIQATraining Dataset=CSIQ2026.030.921—
Re-IQATraining=CSIQ2023.100.919—
Re-IQATraining Dataset=CSIQ2026.030.919—
CONTRIQUETraining=CSIQ2023.100.912—
CONTRIQUETraining Dataset=CSIQ2026.030.912—
HyperIQATraining=KADID2023.100.908—
SHAMISATraining Dataset=TID20132026.030.908—
HyperIQATraining Dataset=KADID-10K2026.030.908—
NIQEType=Handcrafted2026.030.9070.901
CONTRIQUETraining=TID20132023.100.904—
CONTRIQUETraining Dataset=TID20132026.030.904—
ILNIQE2022.040.9020.906
ILNIQEArchitecture Type=CNN/Hand-crafted2023.100.9020.906
SHAMISATraining Dataset=KADID-10K2026.030.901—
Re-IQATraining=TID20132023.100.9—
CONTRIQUETraining=KADID2023.100.9—
Re-IQATraining Dataset=TID20132026.030.9—
CONTRIQUETraining Dataset=KADID-10K2026.030.9—
ARNIQATraining=KADID2023.100.898—
ARNIQATraining Dataset=KADID-10K2026.030.898—
Su et al.Training=KADID2023.100.896—
Su et al.Training Dataset=KADID-10K2026.030.896—
DIIVINE2022.040.8920.908
Su et al.Training=TID20132023.100.892—
Re-IQATraining=KADID2023.100.892—
DIIVINEArchitecture Type=CNN/Hand-crafted2023.100.8920.908
DIIVINE2026.020.8920.908
Su et al.Training Dataset=TID20132026.030.892—
Re-IQATraining Dataset=KADID-10K2026.030.892—
HyperIQATraining=TID20132023.100.876—
HyperIQATraining Dataset=TID20132026.030.876—
ARNIQATraining=TID20132023.100.869—
ARNIQATraining Dataset=TID20132026.030.869—