Sum modified laplacian operator

ML(p, q) =| 2I(p, q) − I(p − 1, q) − I(p + 1, q) | + | 2I(p, q) − I(p, q − 1) − I(p, q + 1) | (2.2.6) where, I(p,q) denotes the pixel value located at the position (p,q).

My Image I is of size 256x256. To implement the above line, I am getting a zero index error as I have choosen p=1:256 and q=1:256. How to avoid getting a zero index error?

Answer 1

[KSSV]

for p = 2:255
    for q = 2:255
        % do what you want 
    end
end

[Psk]

Then the resulting matrix that i will get won't be a 256x256 matrix

[KSSV]

You intialize it as 256*256....for the first and last index (1 and 256) you can use only the which has (p,q) and (p+1,q+1) negelcting p-1, q-1.

Or you resize your 256*256 image into a 258*258 using imresize or interp2.

[Psk]

But in that case that will be a deviation from the original formula. That is if i don't use (p-1) and q-1 for first and last index.

[KSSV]

Then you resize your image into 258*258..