[SciPy-user] cubic splines in scipy

Andrew Straw strawman at astraw.com
Sun Jan 29 22:36:18 CST 2006


Travis Oliphant wrote:

>The specialized cubic-spline algorithms in scipy.signal make use of 
>speed enhancements that are possible when your control points are 
>*equally-spaced* and you are willing to assume specific kinds of 
>boundary conditions (Michael Unser has written extensively about this 
>use of cubic splines and his papers form the basis for what is in 
>scipy.signal).   The big win is that by making these assumptions your 
>matrix-to-be inverted is circulant and the inversion can be accomplished 
>very quickly.
>  
>
Just to let people know, here is a very nice review article by Michael
Unser about splines:
http://bigwww.epfl.ch/publications/unser9902.html

It would be good to include a link to that article within the source
code for those wishing to delve deeper.

Cheers!
Andrew



More information about the SciPy-user mailing list