Table 3 Performance comparison of fast profile and the proposed method
From: Improved BM3D image denoising using SSIM-optimized Wiener filter
Noise | BM3D | Proposed | PSNR | % PSNR | BM3D | Proposed | SSIM | % SSIM |
|---|---|---|---|---|---|---|---|---|
level | PSNR | PSNR | gain | gain | SSIM | SSIM | gain | gain |
10 | 34.18 | 34.17 | –0.01 | –0.029% | 0.904 | 0.903 | –0.001 | –0.111% |
20 | 31.04 | 31.09 | 0.05 | 0.161% | 0.844 | 0.844 | 0.000 | 0.000% |
30 | 29.07 | 29.27 | 0.20 | 0.688% | 0.789 | 0.795 | 0.006 | 0.760% |
40 | 27.45 | 27.91 | 0.46 | 1.676% | 0.732 | 0.752 | 0.020 | 2.732% |
50 | 26.81 | 27.06 | 0.25 | 0.932% | 0.703 | 0.720 | 0.017 | 2.418% |
60 | 25.89 | 26.30 | 0.41 | 1.583% | 0.658 | 0.690 | 0.032 | 4.863% |
70 | 25.07 | 25.63 | 0.56 | 2.234% | 0.614 | 0.663 | 0.049 | 7.980% |
80 | 24.38 | 25.11 | 0.73 | 2.994% | 0.575 | 0.641 | 0.066 | 11.478% |
90 | 23.76 | 24.64 | 0.88 | 3.704% | 0.534 | 0.621 | 0.087 | 16.292% |
100 | 23.14 | 24.17 | 1.03 | 4.451% | 0.500 | 0.600 | 0.100 | 20.000% |