Converting C++ to MATLAB

I am trying to convert some C+开发者_开发问答+ code to MATLAB and having an issue. Would someone please explain a single line of code from below.

  int *image      = (int *) malloc(size*sizeof(int)) ;

  bool diff = true;
  while (diff)
  {
     diff = false;
    for (unsigned int itt = 0; itt < size; itt++)
    {
      diff = diff || (image[itt] != image[image[itt]]);
      image[itt] = image[image[itt]];
    }
  }

Could someone explain from

  diff = diff || (image[itt] != image[image[itt]]);
  image[itt] = image[image[itt]];

This part

image[image[itt]]

I don't understand how this works in terms of MATLAB

Image = zeros(100,100);

ImageSize = (Image);
diff = true;

while(diff)
    diff = false;
    for x=1:ImageSize(1) % Height
        for y=1:ImageSize(2) % Width
        diff = diff || (image(x,y) ~= ????);
        image(x,y) = ????;
        end
    end
end

Thanks to anyone who can help me understand this.

First of all, ImageSize = (Image); is not correct. This will simply assign Image to ImageSize. Use ImageSize = size(Image); instead.

Secondly, you are using double indexing in you Matlab code, while the C++ code is using single indexing.

while(diff)
    diff = false;
    for n=1:numel(Image) %# Total pixels in the image
        diff = diff || (Image(n) ~= Image(Image(n)));
        Image(n) = Image(Image(n));
    end
end

You might have to transpose your matrix, because I'm not certain whether the C++ code operates column-wise or row-wise. Also, because you're initializing you matrix to all zeros (zeros(100,100);), you're going to get all zeros back as the result. Try initializing it to an actual image or at least random integers.

Cheers!