function trainGM()

[f,y] = sampleDetector(@detector1,1000000,8);
fprintf(2,'Fitting model...\n');
save 'samples_gm_1.mat' f y;
beta = logist2(y',f');
save 'beta_gm_1.txt' beta -ascii;

[f,y] = sampleDetector(@detector2,1000000,8);
fprintf(2,'Fitting model...\n');
save 'samples_gm_2.mat' f y;
beta = logist2(y',f');
save 'beta_gm_2.txt' beta -ascii;

[f,y] = sampleDetector(@detector4,1000000,8);
fprintf(2,'Fitting model...\n');
save 'samples_gm_4.mat' f y;
beta = logist2(y',f');
save 'beta_gm_4.txt' beta -ascii;

[f,y] = sampleDetector(@detector8,1000000,16);
fprintf(2,'Fitting model...\n');
save 'samples_gm_8.mat' f y;
beta = logist2(y',f');
save 'beta_gm_8.txt' beta -ascii;

[f,y] = sampleDetector(@detector16,1000000,32);
fprintf(2,'Fitting model...\n');
save 'samples_gm_16.mat' f y;
beta = logist2(y',f');
save 'beta_gm_16.txt' beta -ascii;

function [f] = detector1(im)
f = detector(im,1);

function [f] = detector2(im)
f = detector(im,2);

function [f] = detector4(im)
f = detector(im,4);

function [f] = detector8(im)
f = detector(im,8);

function [f] = detector16(im)
f = detector(im,16);

function [f] = detector(im,sigma)
[m] = detGM(im,sigma); 
m = m(:);
f = [ ones(size(m)) m ]';