Image Quality Assessment on TID 2013 (test)
0.966Mean SRCCMQAF-R
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MQAF-RType=FR, Backbone=ResNet502026.02 | 0.966 | — | — | — | — | — | — | 0.964 | |
| MQAF-RType=FR, Backbone=VGG162026.02 | 0.964 | — | — | — | — | — | — | 0.965 | |
| Bianco et al.2017.09 | 0.96 | 0.96 | — | — | — | — | — | — | |
| MQAF-NRType=NR, Backbone=ResNet502026.02 | 0.96 | — | — | — | — | — | — | 0.963 | |
| TOPIQ-FRType=FR2026.02 | 0.954 | — | — | — | — | — | — | 0.958 | |
| MQAF-NRType=NR, Backbone=VGG162026.02 | 0.951 | — | — | — | — | — | — | 0.953 | |
| Xu et al.2017.09 | 0.95 | 0.96 | — | — | — | — | — | — | |
| JND-SalCARType=FR2026.02 | 0.949 | — | — | — | — | — | — | 0.956 | |
| PieAPPType=FR2026.02 | 0.945 | — | — | — | — | — | — | 0.946 | |
| NIMAArchitecture=VGG162017.09 | 0.944 | 0.941 | 0.538 | 0.557 | 0.054 | — | — | — | |
| DRF-IQA2021.10 | 0.944 | — | — | — | — | — | — | 0.942 | |
| WaDIQaM-FRCategory=FR2022.04 | 0.94 | — | — | — | — | — | — | 0.946 | |
| Ours(JSPL)Category=FR2022.04 | 0.94 | — | — | — | — | — | — | 0.949 | |
| WaDIQaM-FRType=FR2026.02 | 0.94 | — | — | — | — | — | — | 0.946 | |
| DeepQAType=FR2026.02 | 0.939 | — | — | — | — | — | — | 0.947 | |
| Random ForestModel Type=Machine Learning (Ensemble)2026.03 | 0.939 | — | — | — | — | — | — | — | |
| LPIPS-VGGType=FR2026.02 | 0.936 | — | — | — | — | — | — | 0.944 | |
| KGANetType=NR2026.02 | 0.927 | — | — | — | — | — | — | 0.933 | |
| Ours(SL)Category=FR2022.04 | 0.924 | — | — | — | — | — | — | 0.912 | |
| CONTRIQUE-FR2021.10 | 0.909 | — | — | — | — | — | — | 0.915 | |
| VSI2021.10 | 0.902 | — | — | — | — | — | — | 0.903 | |
| VSICategory=FR2022.04 | 0.897 | — | — | — | — | — | — | 0.9 | |
| VSIType=FR2026.02 | 0.894 | — | — | — | — | — | — | 0.898 | |
| SS-IQAType=NR2026.02 | 0.891 | — | — | — | — | — | — | 0.91 | |
| Mittal et al.2017.09 | 0.89 | 0.92 | — | — | — | — | — | — | |
| DB-CNNModel Type=Deep Learning (Black-Box)2026.03 | 0.89 | — | — | — | — | — | — | — | |
| Evo-IQA FullModel Type=Symbolic (White-Box)2026.03 | 0.8898 | — | — | — | — | — | — | — | |
| Moorthy et al.2017.09 | 0.88 | 0.89 | — | — | — | — | — | — | |
| Saad et al.2017.09 | 0.88 | 0.91 | — | — | — | — | — | — | |
| Kottayil et al.2017.09 | 0.88 | 0.89 | — | — | — | — | — | — | |
| MEONModel Type=Deep Learning (Black-Box)2026.03 | 0.88 | — | — | — | — | — | — | — | |
| SSHMPQAEvaluation=Cross-database generalization2026.04 | 0.879 | — | — | — | — | — | — | 0.897 | |
| PieAPP2021.10 | 0.877 | — | — | — | — | — | — | 0.85 | |
| PieAPPEvaluation=Cross-database generalization2026.04 | 0.876 | — | — | — | — | — | — | 0.859 | |
| DeepWSDType=FR2026.02 | 0.874 | — | — | — | — | — | — | 0.87 | |
| HaarPSIModel Type=Manual (Traditional)2026.03 | 0.8662 | — | — | — | — | — | — | — | |
| MSDSEvaluation=Cross-database generalization2026.04 | 0.864 | — | — | — | — | — | — | 0.884 | |
| CaHDCCategory=NR2022.04 | 0.862 | — | — | — | — | — | — | 0.878 | |
| UNIQUE2018.10 | 0.86 | 0.868 | — | — | — | 0.641 | 0.61 | — | |
| DIQaM-FRCategory=FR2022.04 | 0.859 | — | — | — | — | — | — | 0.88 | |
| DeepFL-IQAType=FR2026.02 | 0.858 | — | — | — | — | — | — | 0.876 | |
| MetaIQAType=NR2026.02 | 0.856 | — | — | — | — | — | — | 0.868 | |
| PerSIM2018.10 | 0.853 | 0.854 | — | — | — | 0.655 | 0.64 | — | |
| DISTS2021.10 | 0.853 | — | — | — | — | — | — | 0.873 | |
| FSIM2021.10 | 0.852 | — | — | — | — | — | — | 0.875 | |
| FSIMc2018.10 | 0.851 | 0.832 | — | — | — | 0.727 | 0.68 | — | |
| FSIMCCategory=FR2022.04 | 0.851 | — | — | — | — | — | — | 0.877 | |
| FSIMType=FR2026.02 | 0.851 | — | — | — | — | — | — | 0.876 | |
| PSNR HA2018.10 | 0.847 | 0.85 | — | — | — | 0.615 | 0.65 | — | |
| DeepSimCategory=FR2022.04 | 0.846 | — | — | — | — | — | — | 0.872 | |
| TIQAType=NR2026.02 | 0.846 | — | — | — | — | — | — | 0.858 | |
| CONTRIQUEModel Type=Unsupervised pretraining and Linear Regression2021.10 | 0.843 | — | — | — | — | — | — | 0.857 | |
| HyperIQAModel Type=Supervised pretraining and supervised fine-tuning2021.10 | 0.84 | — | — | — | — | — | — | 0.858 | |
| DIQaM2019.07 | 0.835 | — | — | — | — | — | — | 0.855 | |
| DIQaM-NRCategory=NR2022.04 | 0.835 | — | — | — | — | — | — | 0.855 | |
| A-DISTSType=FR2026.02 | 0.835 | — | — | — | — | — | — | 0.858 | |
| WaDIQaM-NRType=NR2026.02 | 0.835 | — | — | — | — | — | — | 0.855 | |
| BIQA, M.DType=NR2026.02 | 0.835 | — | — | — | — | — | — | 0.859 | |
| DeepIQAEvaluation=Cross-database generalization2026.04 | 0.831 | — | — | — | — | — | — | 0.834 | |
| DISTSCategory=FR2022.04 | 0.83 | — | — | — | — | — | — | 0.855 | |
| DISTSType=FR2026.02 | 0.83 | — | — | — | — | — | — | 0.855 | |
| DISTSEvaluation=Cross-database generalization2026.04 | 0.83 | — | — | — | — | — | — | 0.855 | |
| DEFNetZero-shot=true2025.07 | 0.828 | — | — | — | — | — | — | — | |
| PSNR HMA2018.10 | 0.817 | 0.827 | — | — | — | 0.67 | 0.69 | — | |
| DB-CNN2019.07 | 0.816 | — | — | — | — | — | — | 0.865 | |
| DB-CNNCategory=NR2022.04 | 0.816 | — | — | — | — | — | — | 0.865 | |
| DB-CNNModel Type=Supervised pretraining and supervised fine-tuning2021.10 | 0.816 | — | — | — | — | — | — | 0.865 | |
| DBCNNType=NR2026.02 | 0.816 | — | — | — | — | — | — | 0.865 | |
| SVRModel Type=Machine Learning (Black-Box)2026.03 | 0.8131 | — | — | — | — | — | — | — | |
| SR SIM2018.10 | 0.807 | 0.866 | — | — | — | 0.632 | 0.61 | — | |
| GMSDType=FR2026.02 | 0.804 | — | — | — | — | — | — | 0.858 | |
| FSIMModel Type=Manual (Traditional)2026.03 | 0.8016 | — | — | — | — | — | — | — | |
| Kim et al.2017.09 | 0.8 | 0.8 | — | — | — | — | — | — | |
| IW-CNN2019.07 | 0.8 | — | — | — | — | — | — | 0.802 | |
| IW-CNNCategory=NR2022.04 | 0.8 | — | — | — | — | — | — | 0.802 | |
| NLPDType=FR2026.02 | 0.799 | — | — | — | — | — | — | 0.832 | |
| MS-SSIMModel Type=Manual (Traditional)2026.03 | 0.7915 | — | — | — | — | — | — | — | |
| MS-SSIMCategory=FR2022.04 | 0.786 | — | — | — | — | — | — | 0.83 | |
| MS SSIM2018.10 | 0.785 | 0.83 | — | — | — | 0.691 | 0.69 | — | |
| MADCategory=FR2022.04 | 0.781 | — | — | — | — | — | — | 0.827 | |
| IW SSIM2018.10 | 0.777 | 0.831 | — | — | — | 0.7 | 0.68 | — | |
| MADType=FR2026.02 | 0.773 | — | — | — | — | — | — | 0.803 | |
| UNIQUEZero-shot=true2025.07 | 0.768 | — | — | — | — | — | — | — | |
| WaDIQaM2019.07 | 0.761 | — | — | — | — | — | — | 0.787 | |
| WaDIQaM-NRCategory=NR2022.04 | 0.761 | — | — | — | — | — | — | 0.787 | |
| NIMAArchitecture=Inception-v22017.09 | 0.75 | 0.827 | 0.47 | 0.468 | 0.064 | — | — | — | |
| SSIM2018.10 | 0.741 | 0.788 | — | — | — | 0.733 | 0.76 | — | |
| PQRModel Type=Supervised pretraining and supervised fine-tuning2021.10 | 0.74 | — | — | — | — | — | — | 0.798 | |
| HOSA2019.07 | 0.735 | — | — | — | — | — | — | 0.815 | |
| HOSACategory=NR2022.04 | 0.735 | — | — | — | — | — | — | 0.815 | |
| HOSAModel Type=Codebook-based Features2021.10 | 0.735 | — | — | — | — | — | — | 0.815 | |
| HyperIQACategory=NR2022.04 | 0.729 | — | — | — | — | — | — | 0.775 | |
| MS-SSIMType=FR2026.02 | 0.729 | — | — | — | — | — | — | 0.785 | |
| SSIMCategory=FR2022.04 | 0.727 | — | — | — | — | — | — | 0.777 | |
| BIECON2019.07 | 0.717 | — | — | — | — | — | — | 0.762 | |
| BIECONCategory=NR2022.04 | 0.717 | — | — | — | — | — | — | 0.762 | |
| BIECONModel Type=Supervised pretraining and supervised fine-tuning2021.10 | 0.717 | — | — | — | — | — | — | 0.762 | |
| ResNet-fttraining=fine-tuned2019.07 | 0.712 | — | — | — | — | — | — | 0.756 | |
| ResNet-ftCategory=NR2022.04 | 0.712 | — | — | — | — | — | — | 0.756 | |
| PSNR2018.10 | 0.7 | 0.705 | — | — | — | 0.725 | 0.87 | — |