Laser & Optoelectronics Progress, Volume. 57, Issue 18, 181009(2020)
Super-Resolution Reconstruction of Images Based on Multi-Scale and Multi-Residual Network
Figures & Tables(8)
Fig. 1. Multi-scale and multi-residual super-resolution reconstruction network structure
Fig. 4. Subjective comparison of super-resolution results of butterfly images processed by different algorithms under four times magnification factor. (a) GT; (b) BICUBIC algorithm; (c) ESPCN algorithm; (d) SRCNN algorithm; (e) VDSR algorithm; (f) IMRSR algorithm
Fig. 5. Subjective comparison of super-resolution results of PPT images processed by different algorithms under four times magnification factor. (a) GT; (b) BICUBIC algorithm; (c) ESPCN algorithm; (d) SRCNN algorithm; (e) VDSR algorithm; (f) IMRSR algorithm
Table 1. Effects of number of different multi-scale residual blocks on PSNR
View tableTable 1. Effects of number of different multi-scale residual blocks on PSNR
Number of multi-scale residual block PSNR /dB Set5 Set14 7 31.51 28.18 11 31.52 28.14 15 18 22 31.46 31.59 31.40 28.16 28.19 28.06
Table 2. Average PSNR of different algorithms for different super-resolution reconstructions on different test setsunit: dB
View tableTable 2. Average PSNR of different algorithms for different super-resolution reconstructions on different test setsunit: dB
Dataset Scale BICUBIC Self-Ex SRCNN ESPCN VDSR IMRSR Set5 ×2 ×3 ×4 33.64 30.38 28.42 36.49 32.58 30.31 36.45 32.38 30.19 36.57 32.55 30.31 37.30 33.44 30.98 37.78 33.91 31.59 Set14 ×2 ×3 ×4 30.08 27.38 25.86 32.22 29.16 27.40 32.37 29.15 27.37 32.47 29.27 27.48 32.97 29.69 27.83 33.26 29.88 28.19 BSD100 ×2 ×3 ×4 29.59 27.20 25.96 31.18 28.29 26.84 31.25 28.24 26.80 31.29 28.32 26.85 31.77 28.70 27.14 32.00 28.80 27.30 Urban100 ×2 ×3 ×4 26.86 24.44 23.13 29.54 26.44 24.79 29.08 25.79 24.20 29.21 25.94 24.28 30.46 26.87 24.95 31.00 27.00 25.15
Table 3. Average SSIM of different algorithms for different super-resolution reconstructions on different test sets
View tableTable 3. Average SSIM of different algorithms for different super-resolution reconstructions on different test sets
Dataset Scale BICUBIC Self-Ex SRCNN ESPCN VDSR IMRSR Set5 ×2 ×3 ×4 0.9362 0.8792 0.8223 0.9537 0.9093 0.8619 0.9574 0.9117 0.8639 0.9585 0.9150 0.8672 0.9625 0.9276 0.8875 0.9643 0.9312 0.8957 Set14 ×2 ×3 ×4 0.8807 0.7925 0.7210 0.9034 0.8196 0.7518 0.9141 0.8344 0.7665 0.9153 0.8369 0.7690 0.9199 0.8456 0.7817 0.9227 0.8488 0.7892 BSD100 ×2 ×3 ×4 0.8575 0.7584 0.6856 0.8855 0.7840 0.7106 0.8974 0.8028 0.7297 0.8979 0.8043 0.7308 0.9041 0.8134 0.7413 0.9073 0.8166 0.7469 Urban100 ×2 ×3 ×4 0.8496 0.7509 0.6739 0.8967 0.8088 0.7174 0.8963 0.8007 0.7249 0.8978 0.8054 0.7287 0.9170 0.8350 0.7614 0.9235 0.8403 0.7714
Tools
Get Citation
Copy Citation Text
Xingyu Chen, Weijin Zhang, Weizhi Sun, Ping'an Ren, Ou Ou. Super-Resolution Reconstruction of Images Based on Multi-Scale and Multi-Residual Network[J]. Laser & Optoelectronics Progress, 2020, 57(18): 181009
Paper Information
Category: Image Processing
Received: Jan. 7, 2020
Accepted: Feb. 10, 2020
Published Online: Sep. 2, 2020
The Author Email: Chen Xingyu (2018020669@stu.cdut.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20