Journal of Applied Optics, Volume. 41, Issue 2, 288(2020)
Image denoising algorithm based on wavelet transform and convolutional neural network
Figures & Tables(11)
Fig. 3. Comparison of denoising performance with different algorithms
Fig. 4. Comparison of denoising performance with different algorithms
Fig. 5. Image denoising results of real noisy images based on proposed algorithm
Fig. 6. Image denoising results of remote sensing image based on proposed algorithm
Table 1. PSNR of experimental results with different algorithms
dB View tableView in ArticleTable 1. PSNR of experimental results with different algorithms
dB Methods starfish airplane monarch parrot bridge flowers Average MF 25.31 24.46 25.37 24.84 23.95 25.64 24.93 ST 24.91 24.07 24.69 24.50 23.84 24.91 24.49 EPLL 28.51 28.60 29.39 28.95 26.47 28.98 28.48 NCSR 28.76 28.45 29.43 28.86 26.29 28.77 28.43 WNNM 21.89 22.15 22.17 21.61 22.08 22.49 22.07 Ours 29.69 28.56 29.53 29.50 26.62 29.69 28.93
Table 2. SSIM of experimental results with different algorithms
View tableView in ArticleTable 2. SSIM of experimental results with different algorithms
Methods starfish airplane monarch parrot bridge flowers Average MF 0.698 0 0.591 2 0.702 7 0.609 0 0.622 3 0.640 3 0.643 9 ST 0.670 0 0.551 1 0.649 9 0.569 7 0.640 9 0.603 5 0.614 1 EPLL 0.851 4 0.863 0 0.902 0 0.852 4 0.761 8 0.833 4 0.844 0 NCSR 0.852 9 0.857 6 0.903 0 0.848 7 0.742 3 0.822 7 0.837 9 WNNM 0.625 5 0.689 3 0.750 6 0.709 1 0.465 0 0.621 8 0.643 6 Ours 0.862 8 0.864 9 0.899 9 0.853 8 0.729 6 0.834 4 0.840 9
Table 3. Average time of experimental results with different algorithms
View tableView in ArticleTable 3. Average time of experimental results with different algorithms
Methods MF ST EPLL NCSR WNNM 本文 Average time/s 2.97 0.22 53.53 313.73 240.56 1.46
Table 4. PSNR of experimental results with different algorithms
dB View tableView in ArticleTable 4. PSNR of experimental results with different algorithms
dB Methods C.man lena peppers woman baby bird Average MF 20.53 21.54 21.61 28.48 21.96 21.87 22.67 ST 19.74 20.09 20.18 20.29 20.52 20.42 20.21 EPLL 26.09 28.61 29.05 27.02 28.65 28.06 27.91 NCSR 26.15 28.91 29.22 27.13 28.93 28.00 28.06 WNNM 21.22 25.96 25.53 22.30 26.40 23.59 24.17 Ours 26.63 29.19 29.59 27.42 29.02 28.36 28.37
Table 5. SSIM of experimental results with different algorithms
View tableView in ArticleTable 5. SSIM of experimental results with different algorithms
Methods C.man lena peppers woman baby bird Average MF 0.334 6 0.325 7 0.316 5 0.416 9 0.342 3 0.409 9 0.357 7 ST 0.309 7 0.273 0 0.261 4 0.371 1 0.284 2 0.342 0 0.306 9 EPLL 0.775 4 0.787 0 0.778 2 0.825 3 0.779 9 0.796 7 0.790 4 NCSR 0.783 4 0.803 5 0.790 7 0.842 9 0.792 5 0.814 3 0.804 6 WNNM 0.655 8 0.747 1 0.736 4 0.733 6 0.722 1 0.700 3 0.715 9 Ours 0.783 4 0.797 9 0.789 1 0.851 7 0.792 4 0.833 7 0.808 0
Tools
Get Citation
Copy Citation Text
Qingjiang CHEN, Xiaohan SHI, Yuzhou CHAI. Image denoising algorithm based on wavelet transform and convolutional neural network[J]. Journal of Applied Optics, 2020, 41(2): 288
Paper Information
Category:
Received: May. 7, 2019
Accepted: --
Published Online: Apr. 23, 2020
The Author Email: SHI Xiaohan (910762960@qq.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20