Created by: laurentes

There is an invalid parametrization in the Gaussian filter of the IP module. A factor 1/2 is missing is the argument of the exponential, which means that the variance given as a parameter of the filter corresponds to the variance/2.

This can be easily fixed, but the main issue is that this class is used by the TanTriggs preprocessing. This means that if the fix is applied, the user SHOULD divide the provided variance by 2 to keep getting the same results. Therefore, I don't know what would be the most suitable option.