Hello everybody,<br><br>I'll have a question regarding the TPS
interpolation in 3-D. I'm reading the paper of Davis which is cited in
the itk Documentation.<br><br>Is there a good explanation why the kernel
matrix G considers only | r | and not r^2 log(r) in 3-D ? I can't find a
proper solution, why this holds.<br>
And there are even both implementations in itk, but I'll get better results with the first one.<br>Furthermore,
how do i ensure that the points further away from the control points
are only affected by the affine transformation?<br>
<br>Thanks a lot for answering.<br><br>Best,<br><br>Kerstin