[Insight-users] translation Vs. rotation transforms

[indian rock]

Hi,

I'm trying to match two series (each containing 30 images). I'm computing
the metric value for each image in one series registered with each image in
the other series. So, in total i have 30x30=900 computations. Then i sort
the metric values and the image pair with the lowest/highest metric value is
assumed to be the matching pair.

The problem is for some body parts, i'm getting better results using a
Translation transform and a MeanSquares image to image metric, as compared
to using a CenteredRigid2Dtrasnform with a NormalizedMutual Information
Histogram Image to Image metric..... You would think it should be the other
way around.
Is this possible or am i doing something wrong?

Thanks,
IR.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/insight-users/attachments/20050930/026d504f/attachment.html