ICA_loglikelihood.m

function [ll] = ICA_loglikelihood(X, model)
% Calculate the log likelihood of X, given an ICA model.
% 
% The model assumes each column of X is independently generated by a
% mixture model with parameters theta.
%
% Arguments
%  X - A DxM matrix, whose every column corresponds to a patch in D
%      dimensions (typically D=64).
%  model - A struct with fields:
%           P - mixing matrix of sources (P: D ind. sources -> D signals)
%           vars - a DxK matrix whose (d,k) element correponsds to the
%                  variance of the k'th component in dimension d.
%           mix - a DxK matrix whose (d,k) element correponsds to the
%                 mixing weight of the k'th component in dimension d.
%

[D, K] = size(model.mix);
S = model.P \ X;

ll = 0;
for i=1:D
    S_i = S(i,:);
    theta_i.means = zeros(K, 1);
    theta_i.covs = reshape(model.vars(i,:), 1, 1, K);
    theta_i.mix = model.mix(i,:)';    
    ll = ll + GMM_loglikelihood(S_i, theta_i);
end