gendatatbl03.m

一种基于压缩感知技术的图像重建程序

% GenDataTbl01: Generates the data that appears in Table 1:
% Comparison of running times on a random problem suite, for BP,OMP,FDR,FAR.

rand('state',4159487393);
randn('state',6504972169);

NArr = [20000, 50000];
nArr = [10000, 20000];
kArr = [500,   1000];

tBP = zeros(size(NArr));
tOMP = zeros(size(NArr));
tFDR = zeros(size(NArr));
tFAR = zeros(size(NArr));
eBP = zeros(size(NArr));
eOMP = zeros(size(NArr));
eFDR = zeros(size(NArr));
eFAR = zeros(size(NArr));

for ii = 1:length(NArr)
    ii
    n = nArr(ii);
    N = NArr(ii);
    k = kArr(ii);
    
    x = SparseVector(N,k,'Gaussian',1);
    % Calculate y using the CS matrix
    y = FastCSOperator(1,n,N,x,1:N,N);

    % Initialize threshold parameters
    delta = n/N;
    rho = k/n;
    S = 10;
    alpha_0 = delta*(1-rho)/S;
    q = min((n-k)/k,0.5);

    % Solve using BP (via PDCO)
    tic
    xBP = SolveBP('FastCSOperator', y, N);
    tBP(ii) = toc;
    eBP(ii) = norm(xBP - x);

    % Solve with OMP
    tic
    [xOMP, iters, activeSet] = SolveOMP('FastCSOperator', y, NArr(ii));
    tOMP(ii) = toc;
    eOMP(ii) = norm(xOMP - x);
    
    % Solve with FDR thresholding
    tic
    [xFDR, iters] = SolveStOMP('FastCSOperator', y, NArr(ii), 'FDR', q, S, 1);
    tFDR(ii) = toc;
    eFDR(ii) = norm(xFDR - x);
    
    % Solve with FAR thresholding
    tic
    [xFAR, iters] = SolveStOMP('FastCSOperator', y, NArr(ii), 'FAR', alpha_0, S, 1);
    tFAR(ii) = toc;
    eFAR(ii) = norm(xFAR - x);
end

save DataTbl01.mat NArr nArr kArr tBP tOMP tFDR tFAR eBP eOMP eFDR eFAR

%
% Copyright (c) 2006. David Donoho, Iddo Drori, and Yaakov Tsaig
%  

%
% Part of SparseLab Version:100
% Created Tuesday March 28, 2006
% This is Copyrighted Material
% For Copying permissions see COPYING.m
% Comments? e-mail [e-mail]
%