orldata.m

NMFs算法（带稀疏度约束的非负稀疏矩阵分解）用于实现基于人脸局部特征的人脸识别

function V = orldata
% orldata - read face image data from orl database
%

global imloadfunc;
    
% Reduces the size of the images (by a factor 0.5) 
% Set to 0 to avoid reducing. Set to 1 to reduce.
reducesize = 1; 
    
% This is where the cbcl face images reside
thepath = '../data/orl-faces/';

% Create the data matrix
if reducesize, V = zeros(46*56,400); 
else V = zeros(92*112,400); 
end

% Step through each subject and each image
fprintf('Reading in the images...\n');
i = 0;
for subj=1:40,
    for imag=1:10,
	i = i+1;
	fname = [thepath 's' num2str(subj) '/' num2str(imag) '.pgm'];
	switch imloadfunc,
	 case 'pgma_read',
	  I = pgma_read(fname);
	 otherwise,
	  I = imread(fname);
	end
	if reducesize,
	    V(:,i) = reshape(imresize(I,0.5,'bilinear'),[46*56, 1]);
	else
	    V(:,i) = reshape(I,[92*112, 1]);
	end
    end
    fprintf('[%d/40]',subj);
end

fprintf('\n');

% Same preprocessing as Stan Li et al
minval = min(V);
V = V - ones(size(V,1),1)*minval;
maxval = max(V);
V = (V*255) ./ (ones(size(V,1),1)*maxval);

% Additionally, this is required to avoid having any exact zeros:
% (divergence objective cannot handle them!)
V = max(V,1e-4);

% Finally, divide by 10000 to avoid too large values for nmfsc algorithm
V = V/10000;

% Done!