fft1d_f.asv

压缩感知中一种测量方法。介绍如何构造测量矩阵。

% fft1d_f: implementation of dense FFT-based sampling operator; 
% also refer to fft1d_f: implementation of the transpose of dense
% FFT-based sampling operator; 
% Written by Thong Do, JHU, 2007
% Revised by Lu Gan, April. 2008

function b = fft1d_f(x, OMEGA, p)
% Input parameters:
% x: Original signal vN-vector
% OMEGA: K/2-vector of frequency indices choosen
% p: vector that randomizes samples. 

% b: K-vector 

    % 1d DFT
    N = length(x);
    if max(p)>1,
        % pre-randomizing using permutation vector
        fx = 1/sqrt(N)*fft(x(p));
    elseif max(p) == 1
        % pre-randomizing using Bernoulli vector
        x = x.*p;
        fx = 1/sqrt(N)*fft(x);
    end
  b = [sqrt(2)*real(fx(OMEGA)); sqrt(2)*imag(fx(OMEGA))];