External Email - Use Caution
Hi gang, I have downloaded head MRIs of 1000 subjects from ADNI, together with their FreeSurfer generated aparc+aseg.mgz files. I'm using them in a 3D deep learning network. These MRIs are 256 x 256 x 256, which is too large for GPU training. Therefore, I'm trying to pre-process the volumes by putting a bounding box around the brain, crop the image, and then resize it to a smaller size (e.g. 64 * 100 * 100). To put the bounding box around the brain, I need the brain mask. I don't have access to the brainmasks of these subjects, so I'm trying to use mri_extract_label to make brainmasks for them. Is there a way to tell mri_extract_label to put together all labels in one volume? Best regards, Arman
Arman Avesta, MD Radiology Resident & PhD Student Yale School of Medicine & Yale Graduate School of Arts and Sciences
Email: arman.avesta@yale.edu
You can use use mri_binarize with --min 0.5 and it will take all labels with value 1 and greater. You can create a bounding box with mri_mask with the -bb option
On 5/4/2021 11:53 AM, Arman Avesta wrote:
External Email - Use Caution
Hi gang,
I have downloaded head MRIs of 1000 subjects from ADNI, together with their FreeSurfer generated aparc+aseg.mgz files. I'm using them in a 3D deep learning network. These MRIs are 256 x 256 x 256, which is too large for GPU training. Therefore, I'm trying to pre-process the volumes by putting a bounding box around the brain, crop the image, and then resize it to a smaller size (e.g. 64 * 100 * 100).
To put the bounding box around the brain, I need the brain mask. I don't have access to the brainmasks of these subjects, so I'm trying to use mri_extract_label to make brainmasks for them.
Is there a way to tell mri_extract_label to put together all labels in one volume?
Best regards,
Arman
*Arman Avesta, MD* Radiology Resident & PhD Student Yale School of Medicine & Yale Graduate School of Arts and Sciences
Email: arman.avesta@yale.edu
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu