Laser & Optoelectronics Progress, Volume. 55, Issue 3, 031004(2018)
Multiplicative Denoising Method Based on Deep Residual Learning
Figures & Tables(11)
Fig. 4. Results of different methods for test images under the same noise level. (a) Noise images; (b) Lee method; (c) Frost method; (d) BM3D method; (e) NL method; (f) CNN method
Fig. 5. Detail comparison of different denoising methods.(a) Lee method; (b) Frost method; (c) BM3D method; (d) NL method; (e) CNN method
Fig. 6. Denoising results of CNN method under different noise levels. (a) σ2=0.02; (b) σ2=0.04;(c) σ2=0.06; (d) σ2=0.08; (e) σ2=0.1
Table 1. PSNR of different methods for test images under the same noise leveldB
View tableTable 1. PSNR of different methods for test images under the same noise leveldB
Image Lee Frost BM3D NL CNN Lena 26.40 25.14 20.88 28.22 30.59 Baboon 23.21 22.35 20.54 22.33 26.06 Barbara 23.57 23.20 21.66 23.16 28.45 Boats 25.40 24.14 20.32 25.50 28.65 Peppers 25.97 24.87 20.87 28.45 30.14 Satellite 25.58 23.62 20.06 25.83 28.38 Average 25.02 23.89 20.72 25.58 28.71
Table 2. SSIM of different methods for test images under the same noise level
View tableTable 2. SSIM of different methods for test images under the same noise level
Image Lee Frost BM3D NL CNN Lena 0.647 0.496 0.401 0.733 0.846 Baboon 0.564 0.483 0.517 0.451 0.756 Barbara 0.600 0.510 0.559 0.587 0.855 Boats 0.634 0.510 0.424 0.647 0.784 Peppers 0.642 0.471 0.416 0.731 0.804 Satellite 0.572 0.425 0.344 0.572 0.719 Average 0.610 0.482 0.444 0.620 0.794
Table 3. PSNR of different methods under different noise levelsdB
View tableTable 3. PSNR of different methods under different noise levelsdB
Noise image Lee Frost BM3D NL CNN σ 2=0.0227.71 27.58 25.00 28.95 32.99 σ 2=0.0426.36 25.08 20.87 28.19 30.66 σ 2=0.0625.37 23.60 18.85 27.53 29.25 σ 2=0.0824.55 22.52 17.51 26.97 28.19 σ 2=0.1024.00 21.62 16.48 26.42 27.18 Average 25.60 24.08 19.74 27.61 29.65
Table 4. SSIM of different methods under different noise levels
View tableTable 4. SSIM of different methods under different noise levels
Noise image Lee Frost BM3D NL CNN σ 2=0.020.736 0.602 0.558 0.779 0.882 σ 2=0.040.643 0.495 0.401 0.730 0.843 σ 2=0.060.583 0.432 0.332 0.692 0.817 σ 2=0.080.540 0.391 0.284 0.662 0.795 σ 2=0.100.501 0.357 0.245 0.633 0.774 Average 0.601 0.455 0.364 0.699 0.822
Tools
Get Citation
Copy Citation Text
Ming Zhang, Xiaoqi Lü, Liang Wu, Dahua Yu. Multiplicative Denoising Method Based on Deep Residual Learning[J]. Laser & Optoelectronics Progress, 2018, 55(3): 031004
Paper Information
Category: Image processing
Received: Sep. 5, 2017
Accepted: --
Published Online: Sep. 10, 2018
The Author Email: Lü Xiaoqi ( lxiaoqi@126.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20