Video Quality Assessment On Msu Sr Qa Dataset

평가 지표

KLCC

PLCC

SROCC

Type

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	KLCC	PLCC	SROCC	Type	Paper Title	Repository
3SSIM	0.16365	0.20138	0.21450	FR	-	-
ClipIQA+	0.69774	0.71808	0.56875	NR	Exploring CLIP for Assessing the Look and Feel of Images	-
MUSIQ trained on KONIQ	0.51897	0.59151	0.64589	NR	MUSIQ: Multi-scale Image Quality Transformer	-
FSIM	0.26942	0.35083	0.34996	FR	FSIM: A Feature Similarity Index for Image Quality Assessment	-
MSE	0.12067	0.09428	0.16441	FR	-	-
PieAPP	0.61945	0.75743	0.75215	FR	PieAPP: Perceptual Image-Error Assessment through Pairwise Preference	-
Linearity (Norm-in-Norm Loss)	0.52172	0.62204	0.64382	NR	Norm-in-Norm Loss with Faster Convergence and Better Performance for Image Quality Assessment	-
MS-SSIM	0.07821	0.16035	0.11017	FR	Multiscale structural similarity for image quality assessment
TOPIQ (IAA)	0.40663	0.51061	0.51687	NR	TOPIQ: A Top-down Approach from Semantics to Distortions for Image Quality Assessment	-
DBCNN	0.55139	0.63971	0.68621	NR	Blind Image Quality Assessment Using A Deep Bilinear Convolutional Neural Network	-
Q-Align (IQA)	0.61677	0.74116	0.75088	NR	Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels	-
LPIPS (Alex)	0.43158	0.52385	0.54461	FR	The Unreasonable Effectiveness of Deep Features as a Perceptual Metric	-
PSNR over Y	0.09998	0.13840	0.12914	FR	-	-
TOPIQ trained on PIPAL	0.42811	0.57564	0.55568	FR	TOPIQ: A Top-down Approach from Semantics to Distortions for Image Quality Assessment	-
LPIPS (VGG)	0.41471	0.52820	0.52868	FR	The Unreasonable Effectiveness of Deep Features as a Perceptual Metric	-
ERQA	0.47785	0.60188	0.59345	FR	ERQA: Edge-Restoration Quality Assessment for Video Super-Resolution	-
AHIQ	0.47674	0.62311	0.60468	FR	-	-
VMAF	0.32283	0.40073	0.43219	FR	-	-
DISTS	0.42320	0.55042	0.53346	FR	Image Quality Assessment: Unifying Structure and Texture Similarity	-
Q-Align (IAA)	0.42211	0.50055	0.51521	NR	Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels	-

0 of 60 row(s) selected.