Comment by ipv6ipv4
3 years ago
This is incorrect. The frequency domain inverse of the Gaussian ends up yielding a division by zero. There is no inverse for the Gaussian.
3 years ago
This is incorrect. The frequency domain inverse of the Gaussian ends up yielding a division by zero. There is no inverse for the Gaussian.
That is mathematically true but not practically. Though indeed the Gaussian kernel has lots of zeros [1], in actuality, (a) the zeros themselves are at points, not regions, and therefore of little consequence, and (b) in practice the noise generated from reamplifying frequencies near these zeros can be minimized via techniques such as Wiener deconvolution [2].
[1] https://en.wikipedia.org/wiki/Window_function#Gaussian_windo...
[2] https://en.wikipedia.org/wiki/Wiener_deconvolution
They didn't claim invertible. The de-gaussinization is a reversible process albeit not invertible. I actually say more in this comment
https://news.ycombinator.com/item?id=35111998