Video Super Resolution On Msu Video Upscalers

LPIPS

PSNR

SSIM

Results

Performance results of various models on this benchmark

Model Name	LPIPS	PSNR	SSIM	Paper Title	Repository
SwinIR-Real-B	0.183	28.86	0.830	SwinIR: Image Restoration Using Swin Transformer	-
ESRGAN	-	27.29	0.936	ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks	-
LGFN	-	27.42	0.939	Local-Global Fusion Network for Video Super-Resolution
VEAI-GCG-5	0.292	31.01	0.859	-	-
ESPCN	-	26.25	0.926	Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network	-
VEAI-ALQ-13	0.206	31.00	0.890	-	-
VEAI-ASD-2	0.218	30.55	0.868	-	-
VEAI-GHQ-5	0.210	30.55	0.869	-	-
RealEsrgan-F	0.185	28.82	0.850	Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data	-
Davinci SupScl	0.369	30.10	0.854	-	-
GP-Lines	0.212	29.01	0.822	-	-
DBVSR	-	27.28	0.937	Deep Blind Video Super-resolution	-
VEAI-AD-2	0.195	31.15	0.898	-	-
BasicVsr++RD	0.334	30.98	0.881	BasicVSR++: Improving Video Super-Resolution with Enhanced Propagation and Alignment	-
iSeeBetter	-	27.42	0.939	iSeeBetter: Spatio-Temporal Video Super Resolution using Recurrent-Generative Back-Projection Networks
VEAI-AAM-10	0.278	30.76	0.838	-	-
BSRGAN	0.177	29.27	0.836	Designing a Practical Degradation Model for Deep Blind Image Super-Resolution	-
SRMD	0.349	30.96	0.852	Learning a Single Convolutional Super-Resolution Network for Multiple Degradations	-
SwinIR-Real-S	0.189	28.55	0.845	SwinIR: Image Restoration Using Swin Transformer	-
RealEsrgan	0.181	29.14	0.855	Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data	-

0 of 48 row(s) selected.