基于非局部低秩矩阵重建的图像插值  

Image interpolation based on non-local low-rank matrix reconstruction

作　　者：张鋆萍 张润青 王贵锦[1] Zhang Junping;Zhang Runqing;Wang Guijin(Dept.of Electronic Engineeing,Tsinghua University,Beijing 100084,China;Unit 96271 of PLA,Luoyang Henan 471600,China)

机构地区：[1]清华大学电子工程系 [2]中国人民解放军96271部队

出　　处：《计算机应用研究》2018年第6期1914-1916,共3页Application Research of Computers

摘　　要：传统图像插值方法往往只考虑了图像局部相邻像素之间的关系进行插值,忽略了图像中广泛存在的非局部自相似性。为了充分利用图像中的这种非局部自相似性以提高插值图像质量,提出了基于图像非局部低秩重建模型的图像插值方法,为低秩重建模型提出了一种基于分解为子问题交替迭代求解的高效求解算法。提出的算法能获得更高的主观与客观重建图像质量,实验表明,相对于Bicubic、SAI等图像插值方法能取得平均1.37d B和0.77 d B的PSNR增益。Traditional image interpolation method always only focuses on relationship between adjacent pixels of partial image to interpolate,ignoring non-local self-similarity for extensively existing image. In order to fully utilize the non-local self-similarity in image to improve quality of interpolated image,this paper proposed a novel image interpolation method based on nonlocal low-rank matrix reconstruction model of image. It also presented efficient iteration solution algorithm for low-rank reconstruction model. The given algorithm could obtain higher objective and subjective reconstructed image quality. Furthermore,experiment shows that compared with Bicubic,SAI,etc.,image interpolation method can generate PSNR gain with average of1. 37 d B and 0. 77 d B.

关 键 词：图像插值 低秩矩阵重建 优化 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]