The classification of WM probably takes the most. This includes a lot of things, the longest is the subcortical segmentation. See recon-all-status.log for a list of times of each operation. doug

On 06/08/2013 03:05 PM, Joshua Lee wrote:

Dear Freesurfers,

From the Wiki Freesurfer surface stream follows something like:

1. Affine registration to Tal.
2. Bias correction and skull stripping
3. Classify WM
4. Cut hemisphere planes
5. Tile WM surface
6. Refine WM surface following GM-WM intensity gradients
7. Nudge (nudge?) WM surface to follow the GM-CSF gradient to generate pial surface
8. Labeling Data Set
9. Stats

May I ask which couple of these steps tend to be the most computationally taxing/longest to complete?

Also, I'm not sure I understand #7. Is the WM surface scaled in some sense to approximate the GM-CSF boundary, or is this just a refinement step like #6 of a surface that already approximates the GM-CSF boundary?

Thanks, I've spent most of my time working with the sub-cortical stream.

Joshua

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer