Only mask elements of some value with conv2

I want to use convolution, or something like convolution, to perform sum over neighboring elements. So normally I could use

 x=zeros(20,20);
 r=randi(1000,size(x));
 mask=[ 0 0 1 1 1 0 0
        0 1 1 1 1 1 0
        1 1 1 1 1 1 1
        1 1 1 1 1 1 1
        1 1 1 1 1 1 1
        0 1 1 1 1 1 0
        0 0 1 1 1 0 0];
 myconv=conv2(r,mask);

However, I only want the mask to add elements with the same value as the reference element (center of the mask). How can this be done? Thanks!

Note: this would be easy if there were only a few possible values, but I want it to work for many possible values in the matrix r.

Answer 1

[Matt J]

How about this?

   [N,N]=size(mask);

   vals=im2col(r,[N,N],'sliding');
   ref=vals((N^2+1)/2, :); %assumes odd-dimensions of mask
   result=mask(:).' * (vals.*bsxfun(@eq,vals, ref  ));

result=reshape(result,size(r)-N+1);

It's a modification of conv2(...'valid'). If you wanted conv2(...,'full'), you would simply have to pre-pad r with zeros.

[Christopher]

This seems to work well, thanks!