资源简介
稀疏信号的重构算法:IST算法,OMP算法,Lasso算法,StOMP算法,TwIST算法
代码片段和文件信息
function [xx_debiasobjectivetimesdebias_startmses]= IST(yAtauvarargin)
%%% IST 迭代阈值方法
% This function solves the convex problem
% arg min_theta = 0.5*|| y - A x ||_2^2 + tau ||x||_1
% using the MM/IST algorithm described in the paper
% M. Figueiredo and R. Nowak “A Bound Optimization Approach
% to Wavelet-based Image Deconvolution“ IEEE International
% Conference on Image Processing - ICIP‘2005 Genoa Italy September 2005.
%
% Copyright (2007): Mario Figueiredo and Robert Nowak
%
% ===== Required inputs =============
%
% y: 1D vector or 2D array (image) of observations
%
% A: if y and x are both 1D vectors A can be a
% k*n (where k is the size of y and n the size of x)
% matrix or a handle to a function that computes
% products of the form A*v for some vector v.
% In any other case (if y and/or x are 2D arrays)
% A has to be passed as a handle to a function which computes
% products of the form A*x; another handle to a function
% AT which computes products of the form A‘*x is also required
% in this case. The size of x is determined as the size
% of the result of applying AT.
%
% tau: usually a non-negative real parameter of the objective
% function (see above). It can also be an array the same
% size as x with non-negative entries; in this case
% the objective function weights differently each element
% of x that is it becomes
% 0.5*|| y - A x ||_2^2 + tau^T * abs(x)
%
% ===== Optional inputs =============
%
%
% ‘AT‘ = function handle for the function that implements
% the multiplication by the conjugate of A when A
% is a function handle. If A is an array AT is ignored.
%
% ‘StopCriterion‘ = type of stopping criterion to use
% 0 = algorithm stops when the relative
% change in the number of non-zero
% components of the estimate falls
% below ‘ToleranceA‘
% 1 = stop when the relative
% change in the objective function
% falls below ‘ToleranceA‘
% 4 = stop when the objective function
% becomes equal or less than toleranceA.
% Default = 3.
%
% ‘ToleranceA‘ = stopping threshold; Default = 0.01
%
% ‘Debias‘ = debiasing option: 1 = yes 0 = no.
% Default = 0.
%
% ‘ToleranceD‘ = stopping threshold for the debiasing phase:
% Default = 0.0001.
% If no debiasing takes place this parameter
% if present is ignored.
%
% ‘MaxiterA‘ = maximum number of iterations allowed in the
% main phase of the algorithm.
% Default = 1000
%
% ‘MiniterA‘ = minimum number of iterations performed in the
% main phase of the algorithm.
% Default = 5
%
% ‘MaxiterD‘ = maximum number of iteration
属性 大小 日期 时间 名称
----------- --------- ---------- ----- ----
文件 23413 2010-05-20 14:28 TwIST.m
文件 12923 2011-05-11 15:21 IST.m
文件 1353 2010-03-08 15:51 OMP1.m
文件 12693 2010-12-28 11:04 SolveLasso.m
文件 5576 2010-04-15 19:15 SolveStOMP.m
----------- --------- ---------- ----- ----
55958 5
相关资源
- 一个二维logistics函数实现的图像加密
- 马岳林数字预失真代码
- FISTA 压缩感知恢复算法
- DG配电网重构
- matlab实现的MINIST数据集的分类识别
- distribution-network-reconfiguration 电力系统
- 牛津大学开发的matconvnet工具包
- EWT 经验小波分解
- emd分解重构信号实现降噪
- 6_MultDGPlacOptimizationImperiCompAlgo
- DC_distribution 自己搭建的直流微电网的
- matlab生成logisitc,tent,切比雪夫,改
- 压缩传感重构算法(DCT-OMP)
- CNN 深度学习的卷积神经网络的MATLAB代
- 分布式电力系统状态估计完整程序
- PL_logdist_or_norm
- BarrelDistortion 两个matlab程序
- Wavelet_OMP
- colorhistgram 获得目标的颜色直方图
- histmatch 直方图匹配
- CS 关于压缩感知重构算法
- a-level-set-inariable-distance 李春明提出的
- MNIST-handwritten-digits 手写数字识别数据
- img_bp 图像信号的压缩感知重构算法
- EUdistance 计算两个矩阵之间的欧氏距离
- ColorHistogram 图像检索——提取颜色特
- image-registration
- EM EM算法Matlab实现。最大期望(EM)算
- SRmatlab W.T.Freeman 经典 基于例子的超分
- colorhist 颜色特征提取
评论
共有 条评论