Derivative theorems

theorem of convolution

theorem of gaussian = Laplacian of Gaussian LoG

• instead of doing 2 convolutions, we can just do 1 to get the same result

• derivative of the gaussian DoG filter extracts useful info