计算机与数字工程2017，Vol.45Issue(9)：1842-1847,6.DOI:10.3969/j.issn.1672-9722.2017.09.032
一种改进的基于KRR的单帧图像超分辨率重建算法
An Improved Super Resolution Reconstruction Algorithm of Single Image Based on KRR
摘要
Abstract
To solve the problem of image quality in image super resolution reconstruction,this paper proposes an improved al?gorithm of single image super resolution reconstruction based on Kernel Ridge Regression(KRR). The algorithm uses KRR estima?tion of high-frequency information of potential high resolution image,then the kernel matching pursuit algorithm and gradient de?scent algorithm are combined to obtain a sparse representation based method for solving KRR. To solve the problem of ringing arti?factsoccurring due to the regularization effect,the regression results are post-processed using a prior model of a generic image class, then gets the final high resolution reconstruction image. Experimental results demonstrate that the method has good reconstruction quality in the condition of keeping the time complexity of medium level.关键词
超分辨率重建/KRR算法/稀疏表示/图像先验模型Key words
super-resolution/KRR algorithm/sparse representation/prior model ofimage分类
信息技术与安全科学引用本文复制引用
史国川,龚连友,曹宇剑..一种改进的基于KRR的单帧图像超分辨率重建算法[J].计算机与数字工程,2017,45(9):1842-1847,6.基金项目
安徽省自然科学基金项目(编号:1608085MF140)资助. （编号:1608085MF140）
