Table 1 Comparison results on MaskCLP

From: Sgrgan: sketch-guided restoration for traditional Chinese landscape paintings

Method

PSNR\(\uparrow\)

SSIM\(\uparrow\)

LPIPS\(\downarrow\)

0–15%

15–30%

30–45%

0–15%

15–30%

30–45%

0–15%

15–30%

30–45%

SI [18]

28.73

24.94

21.83

0.783

0.599

0.396

0.177

0.233

0.343

EC [22]

29.82

26.27

23.34

0.821

0.688

0.498

0.141

0.189

0.289

RFR [9]

30.06

26.44

23.52

0.840

0.708

0.523

0.129

0.171

0.272

PD-GAN [23]

30.24

26.62

23.45

0.838

0.712

0.545

0.125

0.175

0.254

CTSDG [26]

30.41

26.84

23.78

0.846

0.726

0.568

0.118

0.169

0.256

MAT [29]

30.54

26.98

24.04

0.860

0.737

0.581

0.112

0.158

0.239

MISF [10]

30.48

26.96

23.86

0.858

0.735

0.546

0.109

0.163

0.242

SGRGAN

\(\varvec{30.59}\)

\(\varvec{27.06}\)

\(\varvec{24.10}\)

\(\varvec{0.867}\)

\(\varvec{0.741}\)

\(\varvec{0.582}\)

\(\varvec{0.105}\)

\(\varvec{0.155}\)

\(\varvec{0.231}\)

  1. \(\uparrow\) Higher values are better, \(\downarrow\) Lower values are better. *Optimal results are displayed in bold, while suboptimal results are underlined