[Insight-users] 3D image registration memory problem

[Smriti Raghunathan]

Hi,
 I have had the same problem using the MI metric in the past. Then I
rewrote the 3D code to use Normalized Mutual Information metric with
Versor3D transform and the versor 3D optimizer. Now the code just
keeps running for days without converging...it ran only 45 iterations
over a period of 72hrs without running any other application on that
computer. Any suggestions, anyone? Is it really supposed to be this
slow? How much RAM should my comp have so that this code runs at least
a bit faster. Any help will be appreciated.

Thanks
Smriti