No-Reference Image Quality Assessment on CSIQ-IQA benchmark leaderboard · SOTA2 Research

No-Reference Image Quality Assessment on CSIQ-IQA

0.962SRCCARNIQA

Evaluation Results

Method
ARNIQA2026.06	0.962	0.973
ReIQA2026.06	0.947	0.96
DB-CNN2026.06	0.946	0.959
ReIQA-Q2026.06	0.944	0.964
CONTRIQUE2026.06	0.942	0.955
SLIDE-IQA-Q2026.06	0.94	0.954
SLIDE-IQA2026.06	0.925	0.947
HyperIQA2026.06	0.923	0.942
TReS2026.06	0.922	0.942
PaQ-2-PiQ2023.04	0.872	0.901
PQR2026.06	0.872	0.901
Re-IQA2023.04	0.844	0.88
CONTRIQUE2023.04	0.843	0.857
HyperIQA2023.04	0.84	0.858
CONTRIQUE2023.04	0.823	—
DB-CNN2023.04	0.816	0.865
Re-IQA2023.04	0.808	—
Re-IQA2023.04	0.804	0.861
PQR2023.04	0.798	—
ReIQA-C2026.06	0.766	0.824
BRISQUE2026.06	0.746	0.829
Hyper-IQA2023.04	0.744	—
PQR2023.04	0.719	—
ImageNet Pretrained (Supervised)2023.04	0.679	0.729
CORNIA2023.04	0.678	0.768
CORNIA2026.06	0.678	0.776
Re-IQA2023.04	0.658	0.736
SLIDE-IQA-C2026.06	0.653	0.764
BRISQUE2023.04	0.604	0.694