Hello,
I have about 10 subjects that I am trying to obtain cortical segmentations. In some of the cases parts of the temporal lobes are not included in the segmentation due to low intensity (or RF bias).
I tried adding control points as described in the tutorial (http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/ControlPoints). The wm/pial boundaries are worse than before, excluding even more parts of wm, even where I placed the control points.
I added about 30-50 controls points mostly in low intensity wm strands that are not correctly segmented.
I was hoping someone may have some tips or hints on good placement of control points, besides the ones on the tutorial. Trial and error is suggested in the tutorial but considering the time it takes to re-run each subject, it is too time consuming.
Thanks in advance.
Deniz Ozgen.
You want to make sure you are placing the control points only on wm voxels that are < 110 and wm voxels that are at least 1 voxel away from a gm voxel in all directions.
If you'd like, you could send me a snapshot of where you placed the control points and I can give you feedback on that. You also would want to be sure this is not an error with the topology fixer which would be more obvious in the wm.mgz (if there are any islands of voxels that should be connected). Allison
freesurfer@nmr.mgh.harvard.edu
-
Allison Stevens -
Deniz Ozgen