ECG-Kit 1.0

File: <base>/common/DelayedCovMat.m (1,440 bytes)
%% (Internal) Delayed covariance matrix calculation
% Function for calculating covariance matrix in the vicinity of
% peaks with k_hb heartbeats of delay
%   
% Example
% 
%   A = DelayedCovMat(x, peaks, win_size, nsamp, k_hb)
% 
% Arguments:
%      +x: signal matrix
% 
%      +peaks: QRS complexes location
% 
%      +win_size: HALF of the size of the win around each QRS complex to
%        compute te cov matrix.
% 
%      +nsamp: size(x,1)
% 
%      +k_hb: delay in number of heartbeats to calculate the cov mat.
% 
% Output:
%     + A: The computed cov mat.
% 
% See also PiCA
% 
% Author: Mariano Llamedo Soria llamedom@electron.frba.utn.edu.ar
% Version: 0.1 beta
% Last update: 14/5/2014
% Birthdate  : 23/4/2013
% Copyright 2008-2015
function A = DelayedCovMat(x, peaks, win_size, nsamp, k_hb)


    peaks = peaks( (peaks - win_size) > 0  & (peaks + win_size ) <= nsamp );
    aux_seq_idx = 1:length(peaks);
    T0 = cell2mat(arrayfun(@(a)( peaks(a) - win_size:peaks(a) + win_size ), aux_seq_idx(1:end-k_hb), 'UniformOutput', false));
    T1 = cell2mat(arrayfun(@(a)( peaks(a) - win_size:peaks(a) + win_size ), aux_seq_idx(1+k_hb:end), 'UniformOutput', false));
    
    x_T0_centered = bsxfun(@minus,x(T0,:),mean(x(T0,:)));  % Remove mean
    x_T1_centered = bsxfun(@minus,x(T1,:),mean(x(T1,:)));  % Remove mean

    A = 1/length(T0)*(x_T0_centered'*x_T1_centered);

    A = (A+A')/2;