[SciPy-dev] fast gaussian filtering

Zachary Pincus zachary.pincus@yale....
Fri Dec 19 12:51:30 CST 2008

> Yes, for kernels smaller than 10 or so it will be faster. But a  
> Gaussian
> is actually not of finite size. So a finite sized Gaussian is  
> actually a
> misnomer; it should be a truncated Gaussian. The recursion does not
> truncate the Gaussian kernel (but at some point the impulse response
> will be so small that it falls into the range of rounding error).

Aah, cool -- that's good to know. I'll definitely keep your  
implementation handy. It would definitely be great up on the cookbook!


More information about the Scipy-dev mailing list