Table 3 Comparison of restoration result indicators of different methods to simulate damaged murals
From: Virtual restoration method of Kizil Grotto murals based on multimodal controlled diffusion models
EvaluationMetrics | PSNR/dB | SSIM | ||||||
|---|---|---|---|---|---|---|---|---|
\({M}_{k}\)-1 | \({M}_{k}\)-2 | \({M}_{k}\)-3 | Mean | \({M}_{k}\)-1 | \({M}_{k}\)-2 | \({M}_{k}\)-3 | Mean | |
Repaint | 24.31 | 29.75 | 28.36 | 27.47 | 0.3765 | 0.5341 | 0.4871 | 0.4659 |
CTSDG | 28.56 | 26.90 | 27.70 | 27.72 | 0.9674 | 0.3939 | 0.5250 | 0.6288 |
ICT | 24.28 | 29.64 | 28.46 | 27.46 | 0.3730 | 0.5294 | 0.5198 | 0.4741 |
SD1.5 | 29.35 | 30.37 | 27.61 | 29.11 | 0.8312 | 0.7405 | 0.4681 | 0.6799 |
BrushNet | 27.83 | 28.95 | 29.65 | 28.81 | 0.8477 | 0.3761 | 0.3485 | 0.5241 |
MuralNet | 28.92 | 25.46 | 26.78 | 27.05 | 0.9701 | 0.3630 | 0.6219 | 0.6517 |
Ours | 33.51 | 31.51 | 29.41 | 31.48 | 0.9521 | 0.7613 | 0.7106 | 0.8080 |
