Hi Mark,
The "matched half-way transforms" approach sounds great, but with
regard to this issue in general --
> smoothing is inherent in trilinear interpolation, some rather
> ugly discontinuities occur with nearest neighbour and some
> ringing and/or variable smoothing can occur with sinc.
-- what are your thoughts on the spline interpolation discussed here?
http://bigwww.epfl.ch/thevenaz/interpolation/
my very limited understanding of it, is that it should offer a better
compromise of sharpness vs. ringing than either polynomial or
windowed-sinc interpolation. But I'd be interested to hear
your/other's views...
Cheers,
Ged.
P.S. Incidentally, it seems to be the favoured interpolation method in
SPM, not that I'm suggesting that means FSL should use it ;-)
http://www.fil.ion.ucl.ac.uk/spm/software/spm2/improvements.html
|