how can I make 2d auto-correlation and 2d cross-correlation for images dataset

I want to classify an image based on 1. make 2D auto-correlation for a dataset 2. then make 2D cross-correlation for the same dataset with an image to measure the similrity. 3. Ater that find the MSE between 2D auto-correlation and 2D cross-correlation. 4. classify the image to the class get the minimum Mean square error (MSE) if true % code

      %the normalized 2-D cross-correlation
      trainedIamge   = gpuArray(Final)   %Final is binary image
      testImage = gpuArray(Final1)       %Final1 is binary image 
      c1 = normxcorr2(trainedIamge,testImage);
      %figure, surf(c), shading flat
      [ypeak, xpeak] = find(c==max(c(:)))

      c2 = normxcorr2(trainedIamge,trainedIamge);  % I don't know how I make it Auto correlation
     %figure, surf(c), shading flat

end

I used the following code for finding MSE between 2D auto-correlation and 2D cross-correlation but it gave me an error.

if true
  % code
   err = immse(mat2gray(c1),mat2gray(c2))

end

I converted c1 and c2 onto grayscale images because it gave an error about the datatypes. the above code is for one image. Do I have to make a loop for a dataset or there is a specific method?

thanks

Answer 1

[Image Analyst]

Yes you need to have a loop. See the FAQ for code examples: http://matlab.wikia.com/wiki/FAQ#How_can_I_process_a_sequence_of_files.3F

[Suheer Ali]

thanks for your answer. my question here is about finding the mse error between autocorrelation and cross correlation. how can I do it? and using using

if true
   c2 = normxcorr2(trainedIamge,trainedIamge);
end

is it true for auto-correlation?

[Image Analyst]

That is the autocorrelation. But it's normalized. You can use xcorr2(). Then use immse() to get the mean square error.