Optics and Precision Engineering, Volume. 30, Issue 19, 2404(2022)
Polarization computational imaging super-resolution reconstruction with lightweight attention cascading network
Figures & Tables(20)
Fig. 1. Network architecture of lightweight attention cascading network
Fig. 12. Visualized results of
Fig. 13. Visualized results of
Fig. 14. Visualized results of
Table 1. 空间注意力网络对重建结果的影响（4 SR）
Table 2. MSAB模块对重建效果的影响（4 SR）
Table 3. 信息细化块对重建性能的影响（4 SR）
Table 4. 不同路径的重建模块客观效果对比（4 SR）
Table 5. 不同网络深度对重建效果的影响（3 SR）
View tableView in ArticleTable 5. 不同网络深度对重建效果的影响（3 SR）
CAB数量 Parameters/K PSNR/dB SSIM 10 484 41.80 0.960 6 11 519 41.76 0.960 4 12 554 41.82 0.960 6 13 589 41.79 0.960 7 14 624 41.81 0.960 6 15 659 41.83 0.960 8
Table 6. Indicator comparison of different SR algorithms on fully polarization image set
View tableView in ArticleTable 6. Indicator comparison of different SR algorithms on fully polarization image set
Scale Method Parameters
/K
Monument image Bridge image Building image Road image PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM ×2 SRCNN 57 44.92 0.988 5 43.82 0.972 7 43.38 0.974 9 42.92 0.965 7 FSRCNN 12 44.77 0.988 0 43.85 0.972 9 43.39 0.974 9 42.88 0.965 4 MSRN 5930 46.74 0.989 6 45.12 0.977 5 44.63 0.978 7 44.11 0.971 1 AWSRN 1396 46.74 0.989 6 45.12 0.977 6 44.64 0.978 7 44.12 0.971 2 Ours 543 46.74 0.989 6 45.12 0.977 6 44.63 0.978 7 44.12 0.971 2 ×3 SRCNN 57 41.14 0.968 4 41.55 0.957 6 40.97 0.958 3 40.99 0.946 4 FSRCNN 12 41.13 0.968 3 41.82 0.958 9 41.13 0.959 1 38.68 0.948 0 MSRN 6114 43.26 0.975 9 43.37 0.967 9 42.77 0.967 8 42.39 0.957 1 AWSRN 1476 43.31 0.976 1 43.38 0.968 1 42.76 0.967 8 42.39 0.957 2 Ours 554 43.25 0.975 9 43.39 0.968 1 42.78 0.967 9 42.39 0.957 2 ×4 SRCNN 57 37.64 0.939 3 38.45 0.932 8 38.05 0.930 3 38.89 0.919 4 FSRCNN 12 37.69 0.938 5 39.36 0.940 1 38.42 0.932 8 38.96 0.919 3 MSRN 6078 40.26 0.957 7 41.41 0.957 1 40.40 0.950 9 40.68 0.939 7 AWSRN 1587 40.30 0.958 0 41.90 0.959 8 40.34 0.950 6 40.67 0.939 5 Ours 576 40.21 0.957 4 41.92 0.959 6 40.58 0.952 2 40.65 0.939 3
Tools
Get Citation
Copy Citation Text
Jie WANG, Guoming XU, Jian MA, Yong WANG, Yi LI. Polarization computational imaging super-resolution reconstruction with lightweight attention cascading network[J]. Optics and Precision Engineering, 2022, 30(19): 2404
Paper Information
Category: Information Sciences
Received: May. 13, 2022
Accepted: --
Published Online: Oct. 27, 2022
The Author Email: XU Guoming (xgm121@163.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20