face1.m

来自「给予主量分析人脸识别pca matlab 有注释」· M 代码 · 共 57 行

clear all;

numb=64;                         %Retain 64 Principal components
pp1=56;pp2=46;                  % Row Size of FACE Image is 112; Column Size of FACE Image is 92
%for i=1:1:300                    % Read 420 YALE FACE Images
 %   st=num2str(i);  s=double(imread(st,'jpg'));
 %   x(:,i)=reshape(s(1:2:112,1:2:92),pp1*pp2,1);   %Chang FACE Image Into 1*10304 Vector
 %end
k=1;
for i=1:35      % Read 420 YALE FACE Images
    for j=1:6
      t=(i-1)*6+j
      st=num2str(t);
      s=double(imread(st,'jpg'));
      x(:,k)=reshape(s(1:2:112,1:2:92),pp1*pp2,1);k=k+1;      %Chang FACE Image Into 1*10304 Vector
  end
end

for i=36:70      % Read 420 YALE FACE Images
    for j=1:6
      t=(i-1)*6+j
      st=num2str(t);
      s=double(imread(st,'jpg'));
      x(:,k)=reshape(s(1:2:112,1:2:92),pp1*pp2,1);k=k+1;      %Chang FACE Image Into 1*10304 Vector
  end
end
%P=size(x,1);% Column size 112*92 of PCA Input Variable;

N=size(x,1); % Row size 112*92 of PCA Input Variable;

mb=mean(x);B1=x-ones(N,1)*mb;%********subtract mean

%********Find eigenvectors vi of B'B (PxP)
%[V,D] = eig (1/(N-1)*(B1*B1'));   %scale factor gives eigvals correct
                                %magnitude for large cov mat 
                                %(assuming sample cov)
                                %(assuming sample cov)
%********Sort eigenvectors
[V,D] = eig (B1*B1');
eigvalvec = max(D);[seigvals, index] = sort(eigvalvec); % sort goes low to high
Vsort = V(:,[fliplr(index)]);

x1=(Vsort(:,1:200))';%(U(:,1:64,))'; 

figure
for i=1:numb
    subplot(8,8,i)
   
    t=mat2gray(reshape(x1(i,:),pp1,pp2),[min(x1(i,:)) max(x1(i,:))]);
    
    imshow(t);
end