[Insight-users] Deformable Registration

[Michelangelo Paci]

HI,
I'm trying to use fem-based deformable registration in order to track some
nuclei during their movement. I started with 2 volumes (50 x 50 x 50 voxels
).
Until now I've played only with E, RhoC and nu parameters. Other parameters
such as I, A, h, ecc are useful?
I'm using a the mean sq metric, but with 60 or more iterations the energy
function increases. I think it's not good.
How does it work? The registration stops when it reaches the maximum number
of iterations or when it arrives at convergence, even if it hasn't reached
the maximum nuber iterations?
Thanks


Michelangelo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/insight-users/attachments/20060508/7fdb5e97/attachment.html