A NEW MEDIAN FORMULA WITH APPLICATIONS TO PDE BASED DENOISING

被引:3
作者
Li, Yingying [1 ]
Osher, Stanley [1 ]
机构
[1] Univ Calif Los Angeles, Dept Math, Los Angeles, CA 90095 USA
关键词
Convex optimization; l(1) minimization; TV denoising; Bregman iterative method; ALGORITHM;
D O I
暂无
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
We develop a simple algorithm for finding the minimizer of the function E(x)=Sigma(n)(i)=(1)w(i)vertical bar x-a(i)vertical bar+F(x), when the w(i) are nonnegative and F is strictly convex. If F is also different tiable and F' is bijective, we obtain an explicit formula in terms of a median. This enables us to obtain approximate solutions to certain important variational problems arising in image denoising. We also present a generalization with E(x)=J(x)+F(x) for J(x) a convex piecewise differentiable function with a finite number of nondifferentiable points.
引用
收藏
页码:741 / 753
页数:13
相关论文
共 16 条