Regarding use of index to obtain the required code (Target avoiding loops)

Query: Suppose i have a bigger matrix (in my case image) named as I. Now, i make a function as

[r,c,v]=find(I>128);

Now i have got the index of rows and columns in two different arrays. Now, i want to extract all those values from bigger matrix for those corresponding rows and columns and multiply those values with a value (let it be x)..and all other values which are not in [r,c] to be multiplied by (let it be 'y')

Can you suggest something for it, without going through for loop.

What i tried was this:

Given: r= array of rows stored size=[p,1] c=array of columns stored=[p,1] I is the original matrix (image in my case)

Implementation:

h=I(r,c); h2=diag(h); h2=h2*x; p=I-h2; p=p*y;

Well, i did try this but the result was ambiguous. Is there some better way to implement this. I am able to do it through loops but i want to implement through matrices only.

All suggestions are most welcome.