I think the resolution is important. The CA1 subfield is only about 2-2.5
mm thick after all.
If this is a feature request thread now, let me add that it would be nice
if the actual training algorithm was integrated into Freesurfer. There is a
lot of disagreement about the most appropriate subfiel
Hi Josh,
as of now, it uses nu.mgz, which is conformed and bias field corrected
with N3.
Koen, do you think we should add a flag to change this? The algorithm is
pretty resolution-independent, and could certainly benefit from
higher-res data.
Cheers,
/Eugenio
On Tue, 2013-05-28 at 13:34 -0700, Jo
I have read that the hippo-subfield algorithm uses the aseg for
initialization, and then uses the original MRI data. By original, does that
mean it up-samples the conformed 1mm data to .5mm, or does that mean it
takes the original original input (i.e. my 0.7 mm isotropic MPRAGE) and
upsamples it to
