Table 1.
Recent advances in ISRR.
Fig 1.
The structure of MSFE module.
Fig 2.
Structure of the SE attention mechanism.
Fig 3.
Framework of the AFF module.
Fig 4.
Structure of the MSRAFFB.
Fig 5.
Architecture of DRUDN.
Fig 6.
Principle and schematic diagram of the residual module.
Fig 7.
The structure of MSRAFFB-Net (Source from: https://pixabay.com/photos/%E6%A4%8D%E7%89%A9-%E5%8F%B6%E5%AD%90-%E7%B2%89%E8%89%B2-%E6%98%A5%E5%A4%A9-%E8%8A%B1%E6%9C%B5-6222600/).
Table 2.
Performance differences under module ablation.
Fig 8.
Effect of number of MSRAFFB modules on MAE, PSNR, and SSIM.
Fig 9.
Effect of number of branches in MSRAFFB on MAE, PSNR, and SSIM.
Fig 10.
Accuracy difference analysis (scale×2) between proposed method and baselines.
Fig 11.
Accuracy difference analysis (scale×4) between proposed method and baselines.
Table 3.
Comparison of image reconstruction quality of different methods.
Fig 12.
Actual performance with a scale factor of *2 (Picture (a) source from: https://pixabay.com/photos/tiger-head-face-feline-wild-cat-2923186/.
Picture (b) source from: https://unsplash.com/photos/a-clock-on-the-side-of-a-brick-building-8d_Sf0Eik_M).
Fig 13.
Accuracy differences for plant and landscape images.
Table 4.
Differences in actual algorithm performance metrics.
Table 5.
Comparison of MSRAFFB-Net and recent advances.
Fig 14.
Comparison of inference speed and reconstruction quality under different noise levels.
Fig 15.
Image reconstruction at different noise levels.
(Source from: https://pixabay.com/zh/photos/fox-animal-wildlife-red-fox-furry-1883658/).