List your company
ResearchFormerly WizWand

No-Reference Image Quality Assessment on TID 2013

0.956SRCCOurs-S+
0.4360.5710.7060.841Apr 2, 2023Oct 16, 2023Apr 30, 2024Nov 13, 2024May 29, 2025Dec 12, 2025Jun 28, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Ours-S+Backbone=Swin-Tiny, Joint training (FR & NR tasks)=true, Architecture Type=Transformer-based2023.100.9560.958
Ours-SBackbone=Swin-Tiny, Joint training (FR & NR tasks)=false, Architecture Type=Transformer-based2023.100.9530.956
Re-IQAModule=content + quality2023.040.9470.96
DB-CNN2023.040.9460.959
Re-IQAModule=quality-aware2023.040.9440.964
CONTRIQUE2023.040.9420.955
MANIQAArchitecture Type=Transformer-based2023.100.9370.943
HyperIQA2023.040.9230.942
SHAMISAType=SSL+LR2026.030.9040.919
Ours-RBackbone=ResNet50, Joint training (FR & NR tasks)=false2023.100.9020.908
SLIDE-IQA2026.060.8990.911
SLIDE-IQA-Q2026.060.8810.892
ARNIQAType=SSL+LR2026.030.880.901
ARNIQA2026.060.880.901
TRESArchitecture Type=CNN/Hand-crafted2023.100.8630.883
TReSType=Supervised2026.030.8630.883
TReS2026.060.8630.883
P2P-BMArchitecture Type=CNN/Hand-crafted2023.100.8620.856
MetaIQAArchitecture Type=CNN/Hand-crafted2023.100.8560.868
TIQAArchitecture Type=Transformer-based2023.100.8460.858
ReIQA-Q2026.060.8440.88
CONTRIQUEType=SSL+LR2026.030.8430.857
CONTRIQUE2026.060.8430.857
ImageNet Pretrained (Supervised)2023.040.840.848
HyperIQAArchitecture Type=CNN/Hand-crafted2023.100.840.858
HyperIQAType=Supervised2026.030.840.858
HyperIQA2026.060.840.858
WaDIQaMArchitecture Type=CNN/Hand-crafted2023.100.8350.855
MM-IQA2026.040.830.845
DBCNNArchitecture Type=CNN/Hand-crafted2023.100.8160.865
DB-CNNType=Supervised2026.030.8160.865
DB-CNN2026.060.8160.865
Su et al.Type=Supervised2026.030.8150.859
MEONArchitecture Type=CNN/Hand-crafted2023.100.8080.824
Re-IQAType=SSL+LR2026.030.8040.861
ReIQA2026.060.8040.861
SHAMISATraining Dataset=KADID-10K2026.030.779—
Re-IQATraining=KADID2023.100.777—
Re-IQATraining Dataset=KADID-10K2026.030.777—
Re-IQAModule=content-aware2023.040.7660.824
ARNIQATraining=KADID2023.100.76—
ARNIQATraining Dataset=KADID-10K2026.030.76—
BRISQUE2023.040.7460.829
PQR2026.060.740.798
HOSAType=Codebook2026.030.7350.815
SHAMISATraining Dataset=CSIQ2026.030.729—
ARNIQATraining=CSIQ2023.100.721—
ARNIQATraining Dataset=CSIQ2026.030.721—
BIECONArchitecture Type=CNN/Hand-crafted2023.100.7170.762
HyperIQATraining=KADID2023.100.706—
HyperIQATraining Dataset=KADID-10K2026.030.706—
SHAMISATraining Dataset=LIVE2026.030.7—
ARNIQATraining=LIVE2023.100.697—
ARNIQATraining Dataset=LIVE2026.030.697—
Su et al.Training=KADID2023.100.687—
Su et al.Training Dataset=KADID-10K2026.030.687—
CORNIA2023.040.6780.776
CORNIAType=Codebook2026.030.6780.768
CORNIA2026.060.6780.768
SLIDE-IQA-CBackbone=DINOv32026.060.6660.757
ReIQA-CBackbone=MoCov22026.060.6580.736
DIIVINEArchitecture Type=CNN/Hand-crafted2023.100.6430.567
CONTRIQUETraining=LIVE2023.100.64—
CONTRIQUETraining Dataset=LIVE2026.030.64—
NBIQA2026.040.6280.695
GM-LOG-BIQA2026.040.6270.662
BRISQUEArchitecture Type=CNN/Hand-crafted2023.100.6260.571
ME-IQABase VLM=EvoQuality2026.030.6190.643
CONTRIQUETraining=KADID2023.100.612—
CONTRIQUETraining Dataset=KADID-10K2026.030.612—
EvoQualityModel Category=VLM-based, Zero-shot evaluation=true, Evolution Round=Round 22025.090.6110.674
BRISQUEType=Handcrafted2026.030.6040.694
BRISQUE2026.060.6040.694
Re-IQATraining=LIVE2023.100.588—
Re-IQATraining Dataset=LIVE2026.030.588—
EvoQualityModel Category=VLM-based, Zero-shot evaluation=true, Evolution Round=Round 12025.090.5870.624
BMPRI2026.040.5830.692
BaselineBase VLM=EvoQuality2026.030.5810.615
Re-IQATraining=CSIQ2023.100.575—
Re-IQATraining Dataset=CSIQ2026.030.575—
CONTRIQUETraining=CSIQ2023.100.57—
CONTRIQUETraining Dataset=CSIQ2026.030.57—
Su et al.Training=LIVE2023.100.561—
Su et al.Training Dataset=LIVE2026.030.561—
Su et al.Training=CSIQ2023.100.55—
Su et al.Training Dataset=CSIQ2026.030.55—
ENIQA2026.040.5450.596
HyperIQATraining=LIVE2023.100.541—
HyperIQATraining=CSIQ2023.100.541—
HyperIQATraining Dataset=LIVE2026.030.541—
HyperIQATraining Dataset=CSIQ2026.030.541—
ILNIQEArchitecture Type=CNN/Hand-crafted2023.100.5210.648
SSEQ2026.040.520.615
ME-IQABase VLM=Q-Insight2026.030.5080.52
BLIINDS-II2026.040.490.521
DIIVINE2026.040.4870.521
CurveletQA2026.040.4710.56
Q-InsightModel Category=VLM-based, Zero-shot evaluation=true2025.090.4680.483
BaselineBase VLM=Q-Insight2026.030.4630.477
Compare2ScoreModel Type=Non-reasoning2026.030.4560.419