# Image Laplacian incomprehension

I am reading and trying to implement a research paper. It uses openGL to define which part of an image to use in order to reconstitute a big image composited of several images.

In order to do so, I have to generate several masks and multiply them together for each image.

We evaluate the per-pixel "focusness" of the source image using a sharpness operator that is the energy of the image Laplacian applied on a window of 20x20 pixels centered on each evaluated points.

My incomprehension is that: isn't Laplacian filter the convolution matrix:

 0  -1   0   or    -1  -1  -1
-1   4  -1         -1   8  -1
0  -1   0         -1  -1  -1


But it's of size 3 x 3 and not 20 x 20. The fact that I'm not a native english speaker maybe fools me on the words and there is something I don't understand in this sentence...