Performance differences under module ablation.

Image super-resolution reconstruction is one of the important application branches of computer vision in many fields. However, in actual complex environments, images are often subject to various interferences, leading to severe distortion in the reconstructed images. To address this issue, this study innovatively combines multi-scale feature extraction (MSFE) and attention feature fusion (AFF). After optimization, the multi-scale recursive attentional feature fusion block (MSRAFFB) and MSRAFFB Network (MSRAFFB-Net) application algorithms are proposed. Simulation experiments on standard datasets demonstrated that MSRAFFB was crucial for enhancing the overall performance of the algorithm, improving reconstruction quality over the baseline. Additionally, through reasonable network structure design (such as increasing module depth and branch complexity), MSRAFFB-Net effectively reduced reconstruction error and improved perceptual quality. The algorithm exhibited high reconstruction accuracy across different magnification factors. Furthermore, to assess the algorithm’s performance in more realistic and complex environments, the study conducted performance experiments in actual application scenarios. The results indicate that the algorithm can effectively improve the resolution and visual quality of reconstructed images while preserving the original image’s characteristic information. In summary, the proposed algorithm significantly enhances the accuracy and robustness of image reconstruction, holding positive implications for advancing the application of computer vision technology in various real-world scenarios requiring high-resolution image processing.