At the moment, this is still a WIP, but I would like to have some feedback on this. Connected to my topic on neurostars, I need a defacing algorithm, that also takes away the ear.
I think I've established this with this PR:
(1) I've taken the already existing mean_reg2mean.nii.gz file, segmented it with SPM12's NewSegment and created a binary mask from the GM, WM and CSF segmentation.
(2) I've dilated this new mask with scipy's binary_dilation function, 25 times and got this new mask, which for the moment I call headmask.
Does it make sense to include such a mask in pydeface? Should it replace the current facemask? Should I create an alternative option?
At the moment, this is still a WIP, but I would like to have some feedback on this. Connected to my topic on neurostars, I need a defacing algorithm, that also takes away the ear.
I think I've established this with this PR: (1) I've taken the already existing
mean_reg2mean.nii.gz
file, segmented it with SPM12'sNewSegment
and created a binary mask from the GM, WM and CSF segmentation. (2) I've dilated this new mask with scipy'sbinary_dilation function
, 25 times and got this new mask, which for the moment I callheadmask
.Does it make sense to include such a mask in
pydeface
? Should it replace the currentfacemask
? Should I create an alternative option?Thanks for the feedback.