Video Quality Assessment on LIVE-VQC (SRCC, PLCC)
0.884SRCCSoft Ranking (Stage 1)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Soft Ranking (Stage 1)Training Stage=Stage 1, Evaluation Protocol=10-fold cross-validation2025.05 | 0.884 | 0.894 | |
| KSVQE2024.02 | 0.861 | 0.883 | |
| qstRectifiers=Spatial and Temporal2024.02 | 0.86 | 0.88 | |
| DOVERprotocol=supervised2026.04 | 0.86 | 0.875 | |
| DOVEREvaluation Protocol=10-fold cross-validation2025.05 | 0.86 | 0.875 | |
| KVQPre-training Dataset=LSVQ [56]2025.03 | 0.859 | 0.879 | |
| KVQprotocol=supervised2026.04 | 0.859 | 0.879 | |
| StarVQA+Pre-training Dataset=ImageNet [7], LSVQ [56], ...2025.03 | 0.857 | 0.874 | |
| FastVQA2024.02 | 0.853 | 0.873 | |
| DOVER2024.02 | 0.853 | 0.872 | |
| FastVQA2024.02 | 0.849 | 0.862 | |
| Fast-VQAPre-training Dataset=LSVQ [56]2025.03 | 0.849 | 0.862 | |
| FastVQAprotocol=supervised2026.04 | 0.849 | 0.865 | |
| FAST-VQAEvaluation Protocol=10-fold cross-validation2025.05 | 0.849 | 0.865 | |
| SimpleVQAprotocol=supervised2026.04 | 0.845 | 0.859 | |
| SimpleVQAEvaluation Protocol=10-fold cross-validation2025.05 | 0.845 | 0.859 | |
| Doveraesthetic branch=false2024.02 | 0.844 | 0.875 | |
| FasterVQAPre-training Dataset=LSVQ [56]2025.03 | 0.843 | 0.858 | |
| MinimalisticVQAEvaluation Protocol=10-fold cross-validation2025.05 | 0.842 | 0.854 | |
| Li222024.02 | 0.841 | 0.839 | |
| DPC-VQAprotocol=few-shot2026.04 | 0.841 | 0.853 | |
| Soft Ranking (Base)Training Stage=Base, Evaluation Protocol=10-fold cross-validation2025.05 | 0.84 | 0.853 | |
| Li et al.Pre-training Dataset=BID [6], LIVE [9], KonIQ-10k [15], SPAQ [8]2025.03 | 0.834 | 0.842 | |
| BVQAprotocol=supervised2026.04 | 0.834 | 0.842 | |
| qtRectifiers=Temporal2024.02 | 0.833 | 0.851 | |
| DOVERInference Time (Sec)=0.047, Testing Protocol=Cross-dataset Testing2024.02 | 0.832 | 0.855 | |
| PVQ2024.02 | 0.827 | 0.837 | |
| PVQPre-training Dataset=PaQ-2-PiQ [55]2025.03 | 0.827 | 0.837 | |
| QPT V2Evaluation Protocol=Cross-dataset2024.07 | 0.827 | 0.853 | |
| PatchVQEvaluation Protocol=10-fold cross-validation2025.05 | 0.827 | 0.837 | |
| FAST-VQATraining data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.826 | 0.845 | |
| FAST-VQAMethod Category=Supervised VQA/IQA models2026.05 | 0.826 | 0.845 | |
| VQTPre-training Dataset=ImageNet [7], Kinetics-400 [18]2025.03 | 0.824 | 0.836 | |
| FAST-VQA2022.10 | 0.823 | 0.844 | |
| FastVQAInference Time (Sec)=0.045, Testing Protocol=Cross-dataset Testing2024.02 | 0.823 | 0.844 | |
| FastVQAEvaluation Protocol=Cross-dataset2024.07 | 0.823 | 0.844 | |
| DisCoVQAPre-training Dataset=None2025.03 | 0.82 | 0.826 | |
| VQ-Jarvis2026.03 | 0.82 | 0.862 | |
| DOVERTraining data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.817 | 0.84 | |
| DOVERMethod Category=Supervised VQA/IQA models2026.05 | 0.817 | 0.84 | |
| BVQA-TCSVT-20222022.10 | 0.816 | 0.824 | |
| BVQAEvaluation Protocol=Cross-dataset2024.07 | 0.816 | 0.824 | |
| CONTRIQUEprotocol=few-shot2026.04 | 0.815 | 0.822 | |
| FasterVQA2022.10 | 0.813 | 0.837 | |
| CONVIQTprotocol=few-shot2026.04 | 0.808 | 0.817 | |
| VQAThinkerMethod Category=Reasoning or reinforcement-learning based LMM evaluators2026.05 | 0.808 | 0.847 | |
| qstInference Time (Sec)=0.159, Testing Protocol=Cross-dataset Testing2024.02 | 0.806 | 0.844 | |
| FasterVQA-MTAMI=true2022.10 | 0.803 | 0.826 | |
| qtInference Time (Sec)=0.159, Testing Protocol=Cross-dataset Testing2024.02 | 0.803 | 0.837 | |
| LMM-PVQA (Soft Ranking Stage 3)Training data=PLVD-P1/ P2/ P3 (700k), Human Label=No, Setting=Zero-shot2025.05 | 0.801 | 0.836 | |
| VersusQMethod Category=Reasoning or reinforcement-learning based LMM evaluators2026.05 | 0.801 | 0.848 | |
| TLVQMPre-training Dataset=NA (pure handcraft)2025.03 | 0.799 | 0.803 | |
| CSPTprotocol=few-shot2026.04 | 0.799 | 0.819 | |
| TLVQM2024.02 | 0.798 | 0.802 | |
| LMM-PVQA (Soft Ranking Stage 2)Training data=PLVD-P1/ P2 (600k), Human Label=No, Setting=Zero-shot2025.05 | 0.798 | 0.833 | |
| LMM-PVQA (Soft Ranking Stage 1)Training data=PLVD-P1 (500k), Human Label=No, Setting=Zero-shot2025.05 | 0.797 | 0.832 | |
| Full-res Swin-T feat., 32 x 4 frames2022.10 | 0.794 | 0.809 | |
| Li22Inference Time (Sec)=27.632, Testing Protocol=Cross-dataset Testing2024.02 | 0.793 | 0.811 | |
| FasterVQA-MSAMI=true2022.10 | 0.791 | 0.818 | |
| LMM-PVQA (Hard Ranking)Training data=PLVD-P1 (500k), Human Label=No, Setting=Zero-shot2025.05 | 0.791 | 0.82 | |
| VQ-Insight2026.03 | 0.79 | 0.835 | |
| VQ-InsightMethod Category=Reasoning or reinforcement-learning based LMM evaluators2026.05 | 0.79 | 0.835 | |
| FAST-VQA-M2022.10 | 0.788 | 0.81 | |
| GSTVQA2024.02 | 0.788 | 0.796 | |
| VQA2-ScorerMethod Category=Supervised VQA/IQA models2026.05 | 0.785 | 0.83 | |
| BUONA-VISTATraining data=None, Human Label=No, Setting=Zero-shot2025.05 | 0.784 | 0.794 | |
| Modular-VQA2026.03 | 0.783 | 0.825 | |
| Q-AlignTraining data=fused [11, 17, 28, 40, 70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.783 | 0.819 | |
| Q-AlignMethod Category=Supervised VQA/IQA models2026.05 | 0.783 | 0.819 | |
| PVQEvaluation Protocol=Cross-dataset, Patch usage=w/o patch2024.07 | 0.781 | 0.781 | |
| Q-Align2026.03 | 0.777 | 0.813 | |
| VQA^22026.03 | 0.776 | 0.823 | |
| MinimalisticVQA(IX)Training data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.775 | 0.821 | |
| MinimalisticVQAMethod Category=Supervised VQA/IQA models2026.05 | 0.775 | 0.821 | |
| VSFA2024.02 | 0.773 | 0.795 | |
| qsInference Time (Sec)=0.159, Testing Protocol=Cross-dataset Testing2024.02 | 0.773 | 0.823 | |
| VSFAPre-training Dataset=None2025.03 | 0.773 | 0.795 | |
| Q-Alignprotocol=supervised2026.04 | 0.773 | 0.829 | |
| DOVER2026.03 | 0.771 | 0.819 | |
| PVQ w/patch2022.10 | 0.77 | 0.807 | |
| PVQEvaluation Protocol=Cross-dataset, Patch usage=w/ patch2024.07 | 0.77 | 0.807 | |
| Fast-VQA2026.03 | 0.769 | 0.815 | |
| Minimalist-VQA2026.03 | 0.765 | 0.812 | |
| VideoLLaMA3 (Qwen2.5-7B)S2I-Tuned=Yes2025.06 | 0.763 | 0.742 | |
| UCDAprotocol=few-shot2026.04 | 0.762 | 0.77 | |
| SimpleVQAMethod Category=Supervised VQA/IQA models2026.05 | 0.762 | 0.799 | |
| 2BiVQAPre-training Dataset=ImageNet [7], KonIQ-10k [15]2025.03 | 0.761 | 0.832 | |
| MinimalisticVQA(VII)Training data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.757 | 0.813 | |
| RAPIQUE2024.02 | 0.754 | 0.786 | |
| VSFAInference Time (Sec)=11.109, Testing Protocol=Cross-dataset Testing2024.02 | 0.753 | 0.795 | |
| StarVQAPre-training Dataset=LSVQ [56]2025.03 | 0.753 | 0.809 | |
| VIDEVAL2024.02 | 0.752 | 0.751 | |
| VIDEVALPre-training Dataset=NA (pure handcraft)2025.03 | 0.752 | 0.751 | |
| SimpleVQAInference Time (Sec)=0.714, Testing Protocol=Cross-dataset Testing2024.02 | 0.749 | 0.789 | |
| PVQ wo/patch2022.10 | 0.747 | 0.776 | |
| SimpleVQA2024.02 | 0.74 | 0.775 | |
| RAPIQUEprotocol=supervised2026.04 | 0.74 | 0.764 | |
| LLaVA-OneVision (Vicuna-v1.1-7B)S2I-Tuned=Yes2025.06 | 0.738 | 0.752 | |
| qbRectifiers=None2024.02 | 0.737 | 0.786 | |
| VSFA2022.10 | 0.734 | 0.772 |