[Insight-users] Modeltoimageragistration1 example difference in Guide and Code

[fjlb at duke.edu]

Hi,

I was looking through the ModelToImageRegistration1.cxx example on the last
release of itk and found that the code differs than the one in the software
guide. In the one I downloaded, the points are transformed into physical
coordinates and then passed to the point list (this was not done in the older
version). However, in the newer version's (the one on latest itk release)
GetValue() method, the points are not transformed from physical points to
images indexes as they are in the version from the software guide. Can someone
please clarify this for me.

Also, I'm trying to write my own metric and I need to implement the
GetDerivative() method. I have a spatial object and I can extract the point
list from it. Should I transform all the points (in image coordinates) to
physical points and then proceed (like its done in the newer
modeltoimageregistration1.cxx)?

Also, when calculating the derivative in GetDerivative() should I transform the
points to image coordinates before getting the jacobian?

Should I use GetValue() in image coordinates (unlike the newer example)?

If someone can clarify these that would be really helpful.

Thanks,

Federico