多尺度特征信息融合的图像超分辨率重建  

Image Super-Resolution Reconstruction with Multi-Scale Feature Information Fusion

作　　者：刘佳 李巍岷 杨大为[1] LIU Jia;LI Weimin;YANG Dawei(School of Information Science and Engineering,Shenyang Ligong University,Shenyang 110158,China)

机构地区：[1]沈阳理工大学信息科学与工程学院,沈阳110158

出　　处：《微处理机》2025年第2期21-27,共7页Microprocessors

基　　金：辽宁省自然科学基金面上项目(2022-MS-276)。

摘　　要：针对图像超分辨率重建方法存在的图像特征信息利用率不高、特征信息丢失等问题,提出一种多尺度特征信息融合的图像超分辨率重建方法。该方法通过稠密多尺度特征提取块和移动窗口自注意力机制,提取图像不同尺度的深层和全局特征信息,提高图像特征信息利用率;利用通道注意力和稠密多尺度特征提取块,对多阶段提取的特征进行信息融合,增强图像的深层特征信息;网络整体和模块内部利用跳跃连接减少特征信息在提取过程中的信息丢失。在公共测试集上的实验结果表明,与其他主流方法比较,该方法拥有更好的图像超分辨率重建效果。Aiming at the problems of low utilization of image feature information and loss of feature information in existing image super-resolution reconstruction methods,this paper proposes a multi-scale feature information fusion-based image super-resolution reconstruction method.The method extracts deep and global feature information at different scales of the image through dense multi-scale feature extraction blocks and a mobile window self-attention mechanism,thereby improving the utilization of image feature information.By utilizing channel attention and dense multi-scale feature extraction blocks,the method fuses the features extracted at multiple stages to enhance the deep feature information of the image.Skip connections are employed both globally within the network and locally within the modules to reduce the loss of feature information during the extraction process.Experimental results on public test datasets demonstrate that,compared with other mainstream methods,the proposed method achieves superior performance in image super-resolution reconstruction.

关 键 词：图像超分辨率重建 多尺度特征提取 自注意力机制 深度学习 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]