Image Quality Assessment on AGIQA-3K
0.9091SRCCMST-CLIPIQA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MST-CLIPIQA2026.06 | 0.9091 | 0.9282 | |
| MST-CLIPIQA*2026.06 | 0.9085 | 0.9283 | |
| LIQE2026.06 | 0.9009 | 0.922 | |
| MA-AGIQA2026.06 | 0.8939 | 0.9273 | |
| MA-AGIQAType=data-driven, Supervision Setting=Supervised2026.02 | 0.893 | 0.927 | |
| MANIQA2026.06 | 0.8916 | 0.9194 | |
| AGQG +IPCEEvaluation Protocol=Intra-dataset, Incorporation=AGQG module2025.12 | 0.8911 | 0.9284 | |
| PKT-12026.07 | 0.8908 | 0.9205 | |
| PKT-22026.07 | 0.8887 | 0.9194 | |
| LIQE2026.07 | 0.8879 | 0.9174 | |
| AGQG +CLIP-AGIQAEvaluation Protocol=Intra-dataset, Incorporation=AGQG module2025.12 | 0.8864 | 0.9268 | |
| IPCEEvaluation Protocol=Intra-dataset2025.12 | 0.8841 | 0.9246 | |
| CLIP-AGIQA2026.07 | 0.8793 | 0.9127 | |
| ELIQType=data-driven, Supervision Setting=Weak-supervised2026.02 | 0.876 | 0.911 | |
| ViLEvaluation Protocol=Intra-dataset2025.12 | 0.875 | 0.9145 | |
| ViT/B/16Evaluation Protocol=Intra-dataset2025.12 | 0.8659 | 0.9115 | |
| CLIP-AGIQAEvaluation Protocol=Intra-dataset2025.12 | 0.8618 | 0.8978 | |
| AGQG +CLIP-IQAEvaluation Protocol=Intra-dataset, Incorporation=AGQG module2025.12 | 0.8614 | 0.9106 | |
| MANIQAType=data-driven, Supervision Setting=Supervised2026.02 | 0.861 | 0.911 | |
| QualiCLIP+Opinion-Unaware=false2024.03 | 0.86 | 0.903 | |
| MANIQA2026.07 | 0.8566 | 0.901 | |
| AMFF-NetEvaluation Protocol=Intra-dataset2025.12 | 0.8565 | 0.905 | |
| ResNet50Evaluation Protocol=Intra-dataset2025.12 | 0.8552 | 0.9072 | |
| StairIQA2026.06 | 0.8543 | 0.8943 | |
| HyperIQAEvaluation Protocol=Intra-dataset2025.12 | 0.8526 | 0.8975 | |
| SSL-IPFEvaluation Protocol=Intra-dataset2025.12 | 0.8523 | 0.9045 | |
| Q-AlignType=data-driven, Supervision Setting=Supervised2026.02 | 0.852 | 0.881 | |
| HyperIQAType=data-driven, Supervision Setting=Supervised2026.02 | 0.85 | 0.904 | |
| HyperIQA2026.06 | 0.8495 | 0.8923 | |
| StairIQA2026.07 | 0.8471 | 0.8954 | |
| CLIP-IQA+Opinion-Unaware=false2024.03 | 0.844 | 0.894 | |
| CLIP-IQA+Type=data-driven, Supervision Setting=Weak-supervised2026.02 | 0.844 | 0.894 | |
| StairIQAEvaluation Protocol=Intra-dataset2025.12 | 0.8439 | 0.8989 | |
| HyperIQA2026.07 | 0.8432 | 0.8924 | |
| CLIP-IQAEvaluation Protocol=Intra-dataset2025.12 | 0.8426 | 0.8053 | |
| AGQG +TIEREvaluation Protocol=Intra-dataset, Incorporation=AGQG module2025.12 | 0.8421 | 0.8977 | |
| StairIQAType=data-driven, Supervision Setting=Supervised2026.02 | 0.834 | 0.893 | |
| MUSIQ2026.06 | 0.8338 | 0.8698 | |
| MGQAEvaluation Protocol=Intra-dataset2025.12 | 0.8283 | 0.8944 | |
| DBCNNType=data-driven, Supervision Setting=Supervised2026.02 | 0.826 | 0.89 | |
| TIEREvaluation Protocol=Intra-dataset2025.12 | 0.8251 | 0.8821 | |
| MG-IQATraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.824 | 0.841 | |
| MUSIQType=data-driven, Supervision Setting=Supervised2026.02 | 0.82 | 0.865 | |
| LinearityIQA2026.06 | 0.8189 | 0.8309 | |
| MUSIQ2026.07 | 0.8188 | 0.8692 | |
| CONTRIQUEOpinion-Unaware=false2024.03 | 0.817 | 0.879 | |
| CONTRIQUEType=data-driven, Supervision Setting=Weak-supervised2026.02 | 0.817 | 0.879 | |
| VGG16Evaluation Protocol=Intra-dataset2025.12 | 0.8167 | 0.8752 | |
| DBCNNEvaluation Protocol=Intra-dataset2025.12 | 0.8154 | 0.8747 | |
| MST-CLIPIQA*prompts=with2026.06 | 0.8124 | 0.8817 | |
| Re-IQAOpinion-Unaware=false2024.03 | 0.811 | 0.874 | |
| Re-IQAType=data-driven, Supervision Setting=Weak-supervised2026.02 | 0.811 | 0.874 | |
| GRepQOpinion-Unaware=false2024.03 | 0.807 | 0.858 | |
| GRepQ-DType=data-driven, Supervision Setting=Weak-supervised2026.02 | 0.807 | 0.858 | |
| LinearityIQA2026.07 | 0.8043 | 0.8191 | |
| ARNIQAOpinion-Unaware=false2024.03 | 0.803 | 0.881 | |
| ARNIQAType=data-driven, Supervision Setting=Weak-supervised2026.02 | 0.803 | 0.881 | |
| ELIQType=data-driven, Supervision Setting=Label-free2026.02 | 0.801 | 0.827 | |
| VQ-R1Training Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.797 | 0.814 | |
| HyperAlignVenue=-2026.01 | 0.7927 | 0.883 | |
| MST-CLIPIQAprompts=without2026.06 | 0.7895 | 0.8666 | |
| CIA-NetVenue=PR’252026.01 | 0.7797 | 0.8687 | |
| IPCEVenue=CVPRW’242026.01 | 0.7697 | 0.8725 | |
| Q-DeepSight2026.04 | 0.768 | 0.822 | |
| Q-insightTraining Dataset=KonIQ, KADIS2026.01 | 0.766 | 0.816 | |
| Q-insightVenue=NeurIPS’25, Methodology=MLLM-Based Methods2026.01 | 0.766 | 0.816 | |
| Q-InsightTraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.766 | 0.783 | |
| Q-Insight (regression)Method Category=MLLM-based: RL training, Backbone=Qwen2.5-VL-7B, Reproduction Protocol=Reported2026.06 | 0.766 | 0.816 | |
| LIQE2026.06 | 0.7638 | 0.848 | |
| VisualQuality-R1Venue=NeurIPS’25, Methodology=MLLM-Based Methods2026.01 | 0.76 | 0.817 | |
| IP-IQAVenue=ICME’242026.01 | 0.7578 | 0.8544 | |
| VisualQuality-R12026.04 | 0.754 | 0.82 | |
| Q-HawkeyeTraining Dataset=KonIQ2026.01 | 0.752 | 0.807 | |
| Q-HawkeyeMethodology=MLLM-Based Methods2026.01 | 0.752 | 0.807 | |
| AMFF-NetVenue=TBC’242026.01 | 0.7513 | 0.8476 | |
| AMFF-NetOriginal Results=Yes2026.06 | 0.7513 | 0.8476 | |
| Q-Insight2026.04 | 0.749 | 0.81 | |
| DeQA-ScoreTraining Dataset=KonIQ, SPAQ, KADID2026.01 | 0.745 | 0.808 | |
| DeQA-ScoreTraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.74 | 0.758 | |
| Qwen-SFTVenue=arXiv’25, Methodology=MLLM-Based Methods2026.01 | 0.739 | 0.813 | |
| DeQA-ScoreTraining Dataset=KonIQ, SPAQ2026.01 | 0.738 | 0.81 | |
| DeQA-Score2026.04 | 0.738 | 0.79 | |
| CLIP-IQA+Venue=AAAI’23, Methodology=Non-MLLM Deep-Learning2026.01 | 0.737 | 0.739 | |
| DeQA-ScoreTraining Dataset=KonIQ, SPAQ, KADID, PIPAL2026.01 | 0.735 | 0.77 | |
| Q-AlignVenue=ICML’24, Methodology=MLLM-Based Methods2026.01 | 0.735 | 0.772 | |
| Q-AlignMethod Category=MLLM-based: SFT training2026.06 | 0.735 | 0.772 | |
| VisualQuality-R1Training Dataset=KADID, SPAQ2026.01 | 0.733 | 0.8 | |
| MR-IQAMethod Category=MLLM-based: RL training, Backbone=Qwen3-VL-2B, Reproduction Protocol=Controlled reproduction2026.06 | 0.732 | 0.804 | |
| Q-AlignTraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.731 | 0.752 | |
| ImageReward2026.06 | 0.7297 | 0.7847 | |
| DeQA-ScoreVenue=CVPR’25, Methodology=MLLM-Based Methods2026.01 | 0.729 | 0.809 | |
| DeQAMethod Category=MLLM-based: SFT training2026.06 | 0.729 | 0.809 | |
| LIQETraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.719 | 0.741 | |
| VQ-R1 (ranking)Method Category=MLLM-based: RL training, Backbone=Qwen3-VL-2B, Reproduction Protocol=Controlled reproduction2026.06 | 0.718 | 0.744 | |
| TOPIQTraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.713 | 0.731 | |
| MANIQATraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.707 | 0.728 | |
| DBCNNVenue=ICSIPA’19, Methodology=Non-MLLM Deep-Learning2026.01 | 0.701 | 0.73 | |
| Re-IQATraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.7 | 0.719 | |
| MUSIQTraining Dataset=KADID-10k, Evaluation=Zero-shot2026.04 | 0.694 | 0.715 | |
| CLIP-IQA+Method Category=Deep-learning-based2026.06 | 0.685 | 0.736 |