Image Quality Assessment on CSIQ (test)
0.979SRCCMQAF-R
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MQAF-RType=FR, Backbone=ResNet502026.02 | 0.979 | 0.977 | |
| TOPIQ-FRType=FR2026.02 | 0.978 | 0.98 | |
| MQAF-RType=FR, Backbone=VGG162026.02 | 0.978 | 0.978 | |
| Ours(JSPL)Category=FR2022.04 | 0.977 | 0.97 | |
| JND-SalCARType=FR2026.02 | 0.976 | 0.977 | |
| PieAPPType=FR2026.02 | 0.973 | 0.975 | |
| MQAF-NRType=NR, Backbone=ResNet502026.02 | 0.972 | 0.971 | |
| MQAF-NRType=NR, Backbone=VGG162026.02 | 0.971 | 0.969 | |
| LPIPS-VGGType=FR2026.02 | 0.967 | 0.97 | |
| Ours(SL)Category=FR2022.04 | 0.965 | 0.968 | |
| DeepWSDType=FR2026.02 | 0.965 | 0.95 | |
| QAL-IQAType=NR2026.02 | 0.963 | 0.97 | |
| WaDIQaM-FRType=FR2026.02 | 0.962 | 0.967 | |
| DeepQAType=FR2026.02 | 0.961 | 0.965 | |
| SS-IQAType=NR2026.02 | 0.96 | 0.969 | |
| MSDSEvaluation=Cross-database generalization2026.04 | 0.955 | 0.953 | |
| KGANetType=NR2026.02 | 0.954 | 0.963 | |
| MADCategory=FR2022.04 | 0.947 | 0.95 | |
| DB-CNNCategory=NR2022.04 | 0.946 | 0.959 | |
| DISTSCategory=FR2022.04 | 0.946 | 0.946 | |
| DBCNNType=NR2026.02 | 0.946 | 0.959 | |
| SSHMPQAEvaluation=Cross-database generalization2026.04 | 0.945 | 0.945 | |
| VSICategory=FR2022.04 | 0.942 | 0.928 | |
| A-DISTSType=FR2026.02 | 0.941 | 0.947 | |
| DBCNN2023.03 | 0.94 | 0.954 | |
| GMSDType=FR2026.02 | 0.939 | 0.938 | |
| LIQE2023.03 | 0.936 | 0.939 | |
| HyperIQA2023.03 | 0.934 | 0.946 | |
| MADType=FR2026.02 | 0.932 | 0.934 | |
| FSIMCCategory=FR2022.04 | 0.931 | 0.919 | |
| DeepFL-IQAType=FR2026.02 | 0.93 | 0.946 | |
| DISTSType=FR2026.02 | 0.929 | 0.928 | |
| DISTSEvaluation=Cross-database generalization2026.04 | 0.929 | 0.928 | |
| VSIType=FR2026.02 | 0.928 | 0.912 | |
| Ma192023.03 | 0.926 | 0.929 | |
| NLPDType=FR2026.02 | 0.925 | 0.913 | |
| HyperIQACategory=NR2022.04 | 0.923 | 0.942 | |
| FSIMType=FR2026.02 | 0.915 | 0.902 | |
| VIFCategory=FR2022.04 | 0.911 | 0.913 | |
| DeepIQAEvaluation=Cross-database generalization2026.04 | 0.909 | 0.901 | |
| MS-SSIMCategory=FR2022.04 | 0.906 | 0.889 | |
| CaHDCCategory=NR2022.04 | 0.903 | 0.914 | |
| LPIPSCategory=FR2022.04 | 0.903 | 0.927 | |
| BIQA, M.DType=NR2026.02 | 0.903 | 0.925 | |
| TreS2023.03 | 0.902 | 0.923 | |
| UNIQUE2023.03 | 0.902 | 0.921 | |
| VIFType=FR2026.02 | 0.899 | 0.898 | |
| MetaIQAType=NR2026.02 | 0.899 | 0.908 | |
| PieAPPEvaluation=Cross-database generalization2026.04 | 0.892 | 0.877 | |
| PieAPPCategory=FR2022.04 | 0.89 | 0.873 | |
| LPIPSEvaluation=Cross-database generalization2026.04 | 0.881 | 0.899 | |
| MS-SSIMType=FR2026.02 | 0.879 | 0.864 | |
| ResNet-ftCategory=NR2022.04 | 0.876 | 0.905 | |
| SSIMCategory=FR2022.04 | 0.865 | 0.852 | |
| WaDIQaM-NRType=NR2026.02 | 0.852 | 0.844 | |
| CONTRIQUE-FREvaluation=Cross-database generalization2026.04 | 0.843 | 0.855 | |
| FRIQUEECategory=NR2022.04 | 0.835 | 0.874 | |
| SSIMType=FR2026.02 | 0.832 | 0.819 | |
| TIQAType=NR2026.02 | 0.825 | 0.838 | |
| CONTRIQUETraining Dataset=LIVE-IQA2021.10 | 0.823 | — | |
| OursCategory=SFT-based and RL-based MLLMs, Condition=Image-conditioned2025.12 | 0.823 | 0.842 | |
| BIECONCategory=NR2022.04 | 0.815 | 0.823 | |
| IW-CNNCategory=NR2022.04 | 0.812 | 0.791 | |
| PSNRCategory=FR2022.04 | 0.81 | 0.819 | |
| ILNIQE2023.03 | 0.808 | 0.851 | |
| PSNRType=FR2026.02 | 0.801 | 0.781 | |
| M3Category=NR2022.04 | 0.795 | 0.839 | |
| DB-CNNTraining Dataset=LIVE-IQA2021.10 | 0.758 | — | |
| BRISQUECategory=NR2022.04 | 0.746 | 0.829 | |
| HyperIQATraining Dataset=LIVE-IQA2021.10 | 0.744 | — | |
| DeQAModel category=SFT-based and RL-based MLLMs, Input Condition=Image-conditioned2026.01 | 0.744 | 0.787 | |
| DeQACategory=SFT-based and RL-based MLLMs, Condition=Image-conditioned2025.12 | 0.744 | 0.787 | |
| HOSACategory=NR2022.04 | 0.741 | 0.823 | |
| Q-AlignModel category=SFT-based and RL-based MLLMs, Input Condition=Image-conditioned2026.01 | 0.737 | 0.671 | |
| Q-AlignCategory=SFT-based and RL-based MLLMs, Condition=Image-conditioned2025.12 | 0.737 | 0.671 | |
| OursCategory=Caption-Only Conditions, Condition=Caption-Only2025.12 | 0.727 | 0.788 | |
| PQRTraining Dataset=LIVE-IQA2021.10 | 0.719 | — | |
| CLIP-IQA+Model category=Deep-Learning Models, Input Condition=Image-conditioned2026.01 | 0.719 | 0.772 | |
| CLIP-IQA+Category=Deep-Learning Models, Condition=Image-conditioned2025.12 | 0.719 | 0.772 | |
| MUSIQModel category=Deep-Learning Models, Input Condition=Image-conditioned2026.01 | 0.71 | 0.771 | |
| MUSIQCategory=Deep-Learning Models, Condition=Image-conditioned2025.12 | 0.71 | 0.771 | |
| C2ScoreModel category=SFT-based and RL-based MLLMs, Input Condition=Image-conditioned2026.01 | 0.705 | 0.735 | |
| C2ScoreCategory=SFT-based and RL-based MLLMs, Condition=Image-conditioned2025.12 | 0.705 | 0.735 | |
| PaQ2PiQ2023.03 | 0.697 | 0.766 | |
| MUSIQ2023.03 | 0.697 | 0.766 | |
| CORNIACategory=NR2022.04 | 0.678 | 0.776 | |
| Q-Insight-ScoreCategory=Caption-Only Conditions, Condition=Caption-Only2025.12 | 0.677 | 0.752 | |
| NIMAModel category=Deep-Learning Models, Input Condition=Image-conditioned2026.01 | 0.649 | 0.695 | |
| NIMACategory=Deep-Learning Models, Condition=Image-conditioned2025.12 | 0.649 | 0.695 | |
| Self-ConsistencyModel category=SFT-based and RL-based MLLMs, Input Condition=Image-conditioned2026.01 | 0.647 | 0.704 | |
| Q-Insight-ScoreModel category=SFT-based and RL-based MLLMs, Input Condition=Image-conditioned2026.01 | 0.64 | 0.685 | |
| Q-Insight-ScoreCategory=SFT-based and RL-based MLLMs, Condition=Image-conditioned2025.12 | 0.64 | 0.685 | |
| Autoencoder-likeModel category=SFT-based and RL-based MLLMs, Input Condition=Image-conditioned2026.01 | 0.636 | 0.683 | |
| Chain-of-ThoughtModel category=SFT-based and RL-based MLLMs, Input Condition=Image-conditioned2026.01 | 0.634 | 0.687 | |
| NIQE2023.03 | 0.631 | 0.719 | |
| KonCept2023.03 | 0.631 | 0.645 | |
| NIQECategory=Hand-Crafted Models, Condition=Image-conditioned2025.12 | 0.628 | 0.718 | |
| MANIQAModel category=Deep-Learning Models, Input Condition=Image-conditioned2026.01 | 0.627 | 0.623 | |
| MANIQACategory=Deep-Learning Models, Condition=Image-conditioned2025.12 | 0.627 | 0.623 | |
| Self-ConsistencyInput Condition=Text-Only, Text Filtering=Score-related words removed2026.01 | 0.621 | 0.665 |