List your company
ResearchFormerly WizWand

No-Reference Image Quality Assessment on SPAQ (test)

0.924SROCCQ-Tacit†
0.385280.525140.6650.80486Oct 25, 2021Jul 20, 2022Apr 14, 2023Jan 8, 2024Oct 2, 2024Jun 27, 2025Mar 23, 2026
Updated 7mo ago

Evaluation results

MethodLinks
Q-Tacit†Method Category=VLM-based, Training Protocol=Multi-task training set in [25] for GRPO calibration2026.030.9240.913
Q-TacitMethod Category=VLM-based, Training Protocol=Trained on KonIQ-10K2026.030.9230.911
HyperIQAModel Type=Supervised pretraining and supervised fine-tuning2021.100.9160.919
CONTRIQUEModel Type=Unsupervised pretraining and Linear Regression2021.100.9140.919
DB-CNNModel Type=Supervised pretraining and supervised fine-tuning2021.100.9110.915
CONTRIQUEOpinion-Unaware=false2024.030.910.912
Re-IQAOpinion-Unaware=false2024.030.9090.912
Q-Insight†Method Category=VLM-based, Training Protocol=Multi-task training set in [25] for GRPO calibration2026.030.9050.907
ResNet-50Model Type=Supervised pretraining and Linear Regression2021.100.9040.909
ARNIQAOpinion-Unaware=false2024.030.9040.909
GRepQOpinion-Unaware=false2024.030.9020.903
CLIP-IQA+Opinion-Unaware=false2024.030.9010.904
Q-InsightMethod Category=VLM-based, Training Protocol=Trained on KonIQ-10K2026.030.8990.903
DeQA-ScoreMethod Category=VLM-based, Training Protocol=Trained on KonIQ-10K2026.030.8960.895
VisualQuality-R1Method Category=VLM-based, Training Protocol=Trained on KonIQ-10K2026.030.8950.894
Q-AlignMethod Category=VLM-based, Training Protocol=Trained on KonIQ-10K2026.030.8870.886
Qwen-2.5-VL-7BMethod Category=VLM-based, Training Protocol=Trained on KonIQ-10K2026.030.8750.874
CLIP-IQA+Method Category=Non-VLM Deep-learning, Training Protocol=Trained on KonIQ-10K2026.030.8640.866
MUSIQMethod Category=Non-VLM Deep-learning, Training Protocol=Trained on KonIQ-10K2026.030.8630.868
Compare2ScoreMethod Category=VLM-based, Training Protocol=Trained on KonIQ-10K2026.030.860.867
NIMAMethod Category=Non-VLM Deep-learning, Training Protocol=Trained on KonIQ-10K2026.030.8560.838
HOSAModel Type=Codebook-based Features2021.100.8460.852
QualiCLIPOpinion-Unaware=true2024.030.8410.851
BRISQUEModel Type=Traditional/Handcrafted Features2021.100.8090.817
MDFSOpinion-Unaware=true2024.030.8050.812
GRepQ-OUOpinion-Unaware=true2024.030.8020.805
ARNIQA-OUOpinion-Unaware=true2024.030.7880.797
ManIQAMethod Category=Non-VLM Deep-learning, Training Protocol=Trained on KonIQ-10K2026.030.7580.768
CLIP-IQAOpinion-Unaware=true2024.030.7330.734
IL-NIQEOpinion-Unaware=true2024.030.7130.724
CORNIAModel Type=Codebook-based Features2021.100.7090.725
CL-MIOpinion-Unaware=true2024.030.7010.701
NIQEModel Type=Traditional/Handcrafted Features2021.100.70.709
NIQEOpinion-Unaware=true2024.030.70.712
CONTRIQUE-OUOpinion-Unaware=true2024.030.6770.685
NIQEMethod Category=Handcrafted, Training Protocol=Trained on KonIQ-10K2026.030.6640.679
Re-IQA-OUOpinion-Unaware=true2024.030.6130.616
BRISQUEMethod Category=Handcrafted, Training Protocol=Trained on KonIQ-10K2026.030.4060.49