Table 1 The effect of artificial parameter ε n to denoising results (dB)
From: Multiplicative noise removal using primal–dual and reweighted alternating minimization
Test images | NSV | \( \varepsilon_{n} = \frac{1}{n + 2} \) | \( \varepsilon_{n} = \frac{1}{{\left( {n + 2} \right)^{2} }} \) | \( \varepsilon_{n} = \frac{1}{{2^{n + 1} }} \) | \( \varepsilon_{n} = \frac{1}{2} \) | \( \varepsilon_{n} = 0.1 \) | New algorithm |
|---|---|---|---|---|---|---|---|
Shape1 | 1/30 | 13.1487 | 13.2599 | 13.1100 | 12.0367 | 12.5180 | 15.6686 |
1/10 | 12.3661 | 11.6466 | 10.6260 | 12.2160 | 12.1057 | 12.3856 | |
Shape2 | 1/30 | 19.4796 | 18.7226 | 18.8344 | 19.2633 | 18.7180 | 19.9760 |
1/10 | 15.8457 | 14.9985 | 15.6592 | 15.1059 | 15.5034 | 16.0540 | |
Barbara | 1/30 | 12.5857 | 12.2622 | 12.0967 | 12.0016 | 12.1491 | 13.0724 |
1/10 | 10.3475 | 9.4710 | 9.5138 | 9.4252 | 10.0087 | 10.3138 | |
Lena256 | 1/30 | 13.5712 | 13.2305 | 13.2450 | 13.1860 | 13.0905 | 14.0905 |
1/10 | 10.5605 | 10.1580 | 10.1726 | 10.1229 | 10.0934 | 10.9022 | |
Cameraman | 1/30 | 16.4723 | 15.1124 | 16.0535 | 15.7450 | 16.1120 | 16.3157 |
1/10 | 13.1132 | 11.2307 | 11.9135 | 11.8059 | 13.0282 | 13.5304 | |
Phantom | 1/30 | 18.7465 | 17.4713 | 18.2407 | 18.0309 | 18.7136 | 20.3229 |
1/10 | 15.6020 | 14.5858 | 15.0706 | 15.0785 | 15.2035 | 15.6289 | |
Average | 14.3199 | 13.5125 | 13.7113 | 13.6682 | 13.9370 | 14.9384 | |