Scale normalized log

One of the proposals for the research of features in a Gaussian scale space it’s to find extremes of the $2^{nd}$ derivative of the normalized Gaussian function

$$ F(x,y,\sigma) = \sigma^2\nabla^2L(x,y,\sigma) = \sigma^2(\nabla^2G(x,y,\sigma)\ast I(x,y)) $$