Video Quality Assessment on LSVQ (test)
0.907SRCCSoft Ranking (Stage 1)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Soft Ranking (Stage 1)Training Stage=Stage 12025.05 | 0.907 | — | 0.904 | |
| Q-ALIGN (1fps) + FAST-VQATraining Set=LSVQtrain, IQA Pre-training=false, fps=12023.12 | 0.899 | 0.899 | — | |
| KVQGFlops=3532025.03 | 0.896 | — | 0.897 | |
| qstInference Time (Sec)=0.159, Testing Protocol=In-dataset Testing2024.02 | 0.895 | 0.895 | — | |
| Modular-VQA2026.03 | 0.894 | — | 0.891 | |
| VQ-Jarvis2026.03 | 0.893 | — | 0.893 | |
| DOVERInference Time (Sec)=0.047, Testing Protocol=In-dataset Testing2024.02 | 0.888 | 0.889 | — | |
| DOVER2026.03 | 0.888 | — | 0.886 | |
| LMM-PVQA (Soft Ranking Stage 3)Training data=PLVD-P1/ P2/ P3 (700k), Human Label=No, Setting=Zero-shot2025.05 | 0.888 | — | 0.884 | |
| LMM-PVQA (Soft Ranking Stage 2)Training data=PLVD-P1/ P2 (600k), Human Label=No, Setting=Zero-shot2025.05 | 0.887 | — | 0.88 | |
| DOVERTraining Set=LSVQtrain, IQA Pre-training=false, Components=aesthetic branch + FAST-VQA2023.12 | 0.886 | 0.887 | — | |
| KSVQE2024.02 | 0.886 | 0.888 | — | |
| qtInference Time (Sec)=0.159, Testing Protocol=In-dataset Testing2024.02 | 0.886 | 0.883 | — | |
| QPT V2Evaluation Protocol=Intra-dataset2024.07 | 0.886 | — | 0.889 | |
| Q-AlignTraining data=fused [11, 17, 28, 40, 70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.886 | — | 0.884 | |
| LMM-PVQA (Soft Ranking Stage 1)Training data=PLVD-P1 (500k), Human Label=No, Setting=Zero-shot2025.05 | 0.886 | — | 0.88 | |
| Q-AlignMethod Category=Supervised VQA/IQA models2026.05 | 0.886 | — | 0.884 | |
| MinimalisticVQA(IX)Training data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.885 | — | 0.882 | |
| MinimalisticVQAMethod Category=Supervised VQA/IQA models2026.05 | 0.885 | — | 0.882 | |
| VersusQMethod Category=Reasoning or reinforcement-learning based LMM evaluators2026.05 | 0.884 | — | 0.886 | |
| Q-ALIGN (1fps)Training Set=LSVQtrain, IQA Pre-training=false, fps=12023.12 | 0.883 | 0.882 | — | |
| CLIPVQAGFlops=6582025.03 | 0.883 | — | 0.885 | |
| Q-Align2026.03 | 0.883 | — | 0.882 | |
| Q-AlignEvaluation Protocol=Intra-dataset2024.07 | 0.883 | — | 0.882 | |
| LMM-PVQA (Hard Ranking)Training data=PLVD-P1 (500k), Human Label=No, Setting=Zero-shot2025.05 | 0.883 | — | 0.866 | |
| VQAThinkerMethod Category=Reasoning or reinforcement-learning based LMM evaluators2026.05 | 0.883 | — | 0.88 | |
| VQA^22026.03 | 0.882 | — | 0.856 | |
| MinimalisticVQA2025.05 | 0.881 | — | 0.879 | |
| Minimalist-VQA2026.03 | 0.88 | — | 0.872 | |
| FAST-VQA2025.05 | 0.88 | — | 0.88 | |
| Soft Ranking (Base)Training Stage=Base2025.05 | 0.88 | — | 0.864 | |
| FAST-VQATraining data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.88 | — | 0.88 | |
| FAST-VQAMethod Category=Supervised VQA/IQA models2026.05 | 0.88 | — | 0.88 | |
| DOVER2025.05 | 0.878 | — | 0.866 | |
| DOVERTraining data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.878 | — | 0.866 | |
| DOVERMethod Category=Supervised VQA/IQA models2026.05 | 0.878 | — | 0.866 | |
| VQA2-ScorerMethod Category=Supervised VQA/IQA models2026.05 | 0.878 | — | 0.872 | |
| Doveraesthetic branch=false2024.02 | 0.877 | 0.878 | — | |
| FAST-VQATraining Set=LSVQtrain, IQA Pre-training=false2023.12 | 0.876 | 0.877 | — | |
| FastVQA2024.02 | 0.876 | 0.877 | — | |
| FastVQAInference Time (Sec)=0.045, Testing Protocol=In-dataset Testing2024.02 | 0.876 | 0.877 | — | |
| Fast-VQAGFlops=279.12025.03 | 0.876 | — | 0.877 | |
| FastVQAEvaluation Protocol=Intra-dataset2024.07 | 0.876 | — | 0.877 | |
| VQ-Insight2026.03 | 0.875 | — | 0.876 | |
| VQ-InsightMethod Category=Reasoning or reinforcement-learning based LMM evaluators2026.05 | 0.875 | — | 0.876 | |
| Fast-VQA2026.03 | 0.874 | — | 0.878 | |
| Faster-VQAGFlops=69.82025.03 | 0.873 | — | 0.874 | |
| SimpleVQAmulti-scale quality fusion strategy=true2022.04 | 0.867 | 0.861 | — | |
| SimpleVQATraining Set=LSVQtrain, IQA Pre-training=true2023.12 | 0.867 | 0.861 | — | |
| SimpleVQA2024.02 | 0.867 | 0.861 | — | |
| SimpleVQAEvaluation Protocol=Intra-dataset2024.07 | 0.867 | — | 0.861 | |
| SimpleVQA2025.05 | 0.867 | — | 0.861 | |
| SimpleVQAInference Time (Sec)=0.714, Testing Protocol=In-dataset Testing2024.02 | 0.866 | 0.863 | — | |
| SimpleVQAmulti-scale quality fusion strategy=false2022.04 | 0.864 | 0.861 | — | |
| SimpleVQAMethod Category=Supervised VQA/IQA models2026.05 | 0.864 | — | 0.861 | |
| MinimalisticVQA(VII)Training data=LSVQ [70], Human Label=Yes, Setting=Zero-shot2025.05 | 0.861 | — | 0.859 | |
| DisCoVQATraining Set=LSVQtrain, IQA Pre-training=false2023.12 | 0.859 | 0.85 | — | |
| Li et al.2022.04 | 0.852 | 0.854 | — | |
| BVQATraining Set=LSVQtrain, IQA Pre-training=true2023.12 | 0.852 | 0.854 | — | |
| Li22Inference Time (Sec)=27.632, Testing Protocol=In-dataset Testing2024.02 | 0.852 | 0.854 | — | |
| Li et al.GFlops=1125372025.03 | 0.852 | — | 0.855 | |
| BVQAEvaluation Protocol=Intra-dataset2024.07 | 0.852 | — | 0.854 | |
| qbInference Time (Sec)=0.159, Testing Protocol=In-dataset Testing2024.02 | 0.849 | 0.843 | — | |
| qsInference Time (Sec)=0.159, Testing Protocol=In-dataset Testing2024.02 | 0.838 | 0.842 | — | |
| PVQv-patch=true2020.11 | 0.827 | 0.828 | — | |
| PVQ2022.04 | 0.827 | 0.828 | — | |
| PVQTraining Set=LSVQtrain, IQA Pre-training=true2023.12 | 0.827 | 0.828 | — | |
| PVQ2024.02 | 0.827 | 0.828 | — | |
| PVQw/patchGFlops=585012025.03 | 0.827 | — | 0.828 | |
| PVQEvaluation Protocol=Intra-dataset, Patch usage=w/ patch2024.07 | 0.827 | — | 0.828 | |
| PatchVQ2025.05 | 0.827 | — | 0.828 | |
| PVQv-patch=false2020.11 | 0.814 | 0.816 | — | |
| PVQwo/patchGFlops=585012025.03 | 0.814 | — | 0.816 | |
| PVQEvaluation Protocol=Intra-dataset, Patch usage=w/o patch2024.07 | 0.814 | — | 0.816 | |
| VSFA2020.11 | 0.801 | 0.796 | — | |
| VSFA2022.04 | 0.801 | 0.796 | — | |
| VSFATraining Set=LSVQtrain, IQA Pre-training=false2023.12 | 0.801 | 0.796 | — | |
| VSFA2024.02 | 0.801 | 0.796 | — | |
| VSFAInference Time (Sec)=11.109, Testing Protocol=In-dataset Testing2024.02 | 0.801 | 0.796 | — | |
| VSFAGFlops=409192025.03 | 0.801 | — | 0.796 | |
| VSFAEvaluation Protocol=Intra-dataset2024.07 | 0.801 | — | 0.796 | |
| VSFA2025.05 | 0.801 | — | 0.796 | |
| VIDEVAL2024.02 | 0.795 | 0.783 | — | |
| VisualQuality-R1Method Category=Reasoning or reinforcement-learning based LMM evaluators2026.05 | 0.795 | — | 0.796 | |
| VIDEVAL2020.11 | 0.794 | 0.783 | — | |
| VIDEVAL2022.04 | 0.794 | 0.783 | — | |
| VIDEVALTraining Set=LSVQtrain, IQA Pre-training=false2023.12 | 0.794 | 0.783 | — | |
| VIDEVAL2025.03 | 0.794 | — | 0.783 | |
| VIDEVALEvaluation Protocol=Intra-dataset2024.07 | 0.794 | — | 0.783 | |
| VIDEAL2025.05 | 0.794 | — | 0.783 | |
| VideoLLaMA3 (Qwen2.5-7B)S2I-Tuned=Yes2025.06 | 0.793 | — | 0.788 | |
| TLVQM2020.11 | 0.772 | 0.774 | — | |
| TLVQM2022.04 | 0.772 | 0.774 | — | |
| TLVQMTraining Set=LSVQtrain, IQA Pre-training=false2023.12 | 0.772 | 0.774 | — | |
| TLVQM2024.02 | 0.772 | 0.774 | — | |
| TLVQM2025.03 | 0.772 | — | 0.774 | |
| TLVQMEvaluation Protocol=Intra-dataset2024.07 | 0.772 | — | 0.774 | |
| TLVQM2025.05 | 0.772 | — | 0.774 | |
| LLaVA-Video (Qwen2-7B)S2I-Tuned=Yes2025.06 | 0.76 | — | 0.734 | |
| LLaVA-OneVision (Vicuna-v1.1-7B)S2I-Tuned=Yes2025.06 | 0.751 | — | 0.73 |