🚀 go-pugleaf

In article <32FBCF6A.7EF7@radiol.med.uni-giessen.de>,
Max.F.Puille <Max.F.Puille@radiol.med.uni-giessen.de> wrote:
:Hi Jeff,
:I'm afraid this won't be very helpful from a programmer's point of view,
:just a comment from someone doing nuclear medicine, and image coregistration
:in clinical routine for a year or so.
:
:Jeff Tsao wrote:
:...
:Sorry. I'm only 90% sure for MR/CT coregistration, but SPECT/MR or
:SPECT/CT coregistration only can be automatic for one single kind of
:radioactive tracer. For every different tracer you will definitely need
:a different algorithm.
 
Yes, this is largely because each tracer will present a different
nonlinear system to be optimized and, as I mentioned previously, since
nonlinear systems typically are nontypical, if the optimizer Is robust
enough to handle all these cases it likely will have to be tuned for
some to all of these systems.
 
:...
:Now replace 'you' with 'the algorithm' and think about it. Add the fact
:that up to now there is _no_ working algorithm to do automatic segmentation
:of, say, the complete colon in a CT scan. I spoke to H.P. Meinzer at the
:German Cancer Research Center recently, he's on that problem for about
:15 years now. Summary: it doesn't work, only in experimental settings.
:Period.
 
Ditto.
 
:Conclusion: my best wishes for your work regarding MR/CT. But as far as
:NucMed is concerned: do yourself a favor, forget it.
 
I assume this is in jest.  I would not discourage any graduate student
from taking on a hard-to-impossible project of his/her choice that can
lead to some contribution in his/her field.  I am not expert in this
field, but from the postings, the subject seems important enough that
many people are spending lots of resources to come to grips with these
problems.
 
Lester

--
 /*             RESEARCH                            ingber@ingber.com *
  *       INGBER                                 ftp://ftp.ingber.com *
  * LESTER                                     http://www.ingber.com/ *
  * Prof. Lester Ingber __ PO Box 857 __ McLean, VA 22101-0857 __ USA */