1 | initial version |
Why would you want to do that?
Downscale your image and apply a "standard-sized" Gaussian kernel. Reducing the image size amounts to removing the highest frequencies in the image which in turn is equivalent to using a large Gaussian kernel.