Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels

Benchmark Model Rank Results
aesthetics-quality-assessment-on-aestheticOneAlign#1SRCC: 0.823
image-quality-assessment-on-koniq-10kOneAlign#2SRCC: 0.941PLCC: 0.952
video-quality-assessment-on-live-fb-lsvqOneAlign + FAST-VQA#1PLCC: 0.900
video-quality-assessment-on-live-fb-lsvqOneAlign#3PLCC: 0.886
video-quality-assessment-on-msu-sr-qa-datasetQ-Align (IQA)#2SROCC: 0.75088PLCC: 0.74116KLCC: 0.61677Type: NR
video-quality-assessment-on-msu-sr-qa-datasetQ-Align (VQA)#3SROCC: 0.71812PLCC: 0.71121KLCC: 0.58634Type: NR
video-quality-assessment-on-msu-sr-qa-datasetQ-Align (IAA)#34SROCC: 0.51521PLCC: 0.50055KLCC: 0.42211Type: NR