摘要
结构矩阵低秩逼近在图像压缩、计算机代数和语音编码中有广泛应用.首先给出了几类结构矩阵的投影公式,再利用交替投影方法计算结构矩阵低秩逼近问题.数值试验表明新方法是可行的.
Structured low rank approximation has a wide range of applications in image compression,computer algebra,and speech encoding.We first present projection formulas of several kinds of structured matrices and then apply alternating projection algorithm to obtain the solution of structured low rank approximation of a matrix.Finally,the computational results show that the new method is feasible.
出处
《应用数学与计算数学学报》
2014年第4期475-485,共11页
Communication on Applied Mathematics and Computation
基金
国家自然科学基金资助项目(11101100
11261014
11301107)
广西自然科学基金资助项目(2012GXNSFBA053006
2013GXNSFBA019009
2011GXNSFA018138)
广西信息科学实验中心基金资助项目(20130103)
关键词
结构矩阵
低秩逼近
交替投影方法
structured matrix
low rank approximation
alternating projection algorithm
