Why i dont see my image after pca analysis?

Hello, I am working on an image to reduce its size without comprising much on the quality. i gound the first 300 eigenvectors are the most significant and i process my image based on those. below is my coding. when i try to plot it i get only grey color screen. any thought on fixing it will be appreciated highly.

 >> I=imread('abc.jpg');
  >> I2 = im2double(I);
  >> II=I2(:,:,1);
  >> [coeff,score,latent] = pca(II);
  >> Me=mean(II);
  >> plot(0:2847,[0;cumsum(latent)/sum(latent)],'-ob');
  >> plot(0:400,[0;cumsum(latent(1:400))/sum(latent)],'-ob');
  >> C300=coeff(:,300);
  >> S300=score(:,300);
  >> Z300=C300*S300';
  >> Z300=Z300';
  >> MM=repmat(Me,2848,1);
  >> Z300M=Z300+MM;
  >> Z300mu=uint8(Z300M);
  >> image(repmat(Z300mu,1,1,3))

Answer 1

[Image Analyst]

What is the range of Z300mu and what is its data type? Try using imshow() with the [] option, instead of image().

Answer 2

[Chris Jademan]

Hello Hydro, I have just applied the code you had given and there is an error.

Error in CompressImage (line 4) [coeff,score,latent] = pca(II);

What should I do to fix it ?

[Image Analyst]

Give the complete error (ALL the red text), not just the line of code.

Answer 3

[Brendan Hamm]

So the fact that you have an indexing:

II=I2(:,:,1);

implies that II is a m-by-n-by-1 array (3-dimensional array). The pca function expects a m-by-n matrix. You can achieve this with the squeeze function:

II = squeeze(I2(:,:,1));

which will remove the singleton dimension of your 3-d array, then the pca function should work.