Skip to main content

Table 1 Quantitative comparison results of different methods on dataset MSACCSD (“↑” indicates that higher is better. “↓” indicates that lower is better. The best results are marked in bold black)

From: EA-GAN: restoration of text in ancient Chinese books based on an example attention generative adversarial network

Method

PSNR↑

SSIM↑

LPIPS(AlexNet)↓

LPIPS(VGG)↓

GLCIC

23.15

0.905

0.0355

0.0589

PConv

24.19

0.919

0.0316

0.0567

GatedConv

22.98

0.907

0.0418

0.0617

EdgeConnect

25.45

0.924

0.0313

0.0617

RFR-net

25.24

0.933

0.0254

0.0428

Ours

27.95

0.950

0.0165

0.0358