Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data

22 Jul 2021  ·  Xintao Wang, Liangbin Xie, Chao Dong, Ying Shan ·

Though many attempts have been made in blind super-resolution to restore low-resolution images with unknown and complex degradations, they are still far from addressing general real-world degraded images. In this work, we extend the powerful ESRGAN to a practical restoration application (namely, Real-ESRGAN), which is trained with pure synthetic data... Specifically, a high-order degradation modeling process is introduced to better simulate complex real-world degradations. We also consider the common ringing and overshoot artifacts in the synthesis process. In addition, we employ a U-Net discriminator with spectral normalization to increase discriminator capability and stabilize the training dynamics. Extensive comparisons have shown its superior visual performance than prior works on various real datasets. We also provide efficient implementations to synthesize training pairs on the fly. read more

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Video Super-Resolution MSU Video Super Resolution Benchmark Real-ESRGAN Subjective score 3.647 # 11
ERQAv1.0 0.663 # 13
QRCRv1.0 0.000 # 16
SSIM 0.774 # 21
CRRMv1.0 0.942 # 21
PSNR 24.441 # 20
FPS 0.991 # 4
Video Super-Resolution MSU Video Super Resolution Benchmark Real-ESRNet Subjective score 1.715 # 21
ERQAv1.0 0.598 # 21
QRCRv1.0 0.000 # 16
SSIM 0.824 # 16
CRRMv1.0 0.951 # 19
PSNR 27.195 # 12
FPS 0.982 # 3

Methods