Skip to main content

Table 4 The average value of various types of indicators achieved by different algorithms on the images of the mural data set

From: Superresolution reconstruction method for ancient murals based on the stable enhanced generative adversarial network

  PSNR SSIM FID IS MS-SSIM NIQE
RDB + Calculate the perceptual loss after activation + No WGAN-GP 28.59 0.751 17.1 17.9 0.768 7.521
RDB + Calculate the perceptual loss before activation + No WGAN-GP 29.93 0.762 15.1 18.5 0.770 4.621
RDB + Calculate the perceptual loss before activation + WGAN-GP 29.99 0.769 14.8 18.7 0.771 4.623
RS-RDB + Calculate the perceptual loss after activation + No WGAN-GP 30.45 0.784 12.34 19.6 0.774 6.497
RS-RDB + Calculate the perceptual loss before activation + No WGAN-GP 32.41 0.801 11.6 20.5 0.801 4.016
RS-RDB + Calculate the perceptual loss before activation + WGAN-GP 32.53 0.816 11.5 20.8 0.805 3.969