user-forum > mask files of each anatomical structure
Showing 1-4 of 4 posts
Display:
Results per page:
Sep 13, 2022  03:09 PM | ymmpi
mask files of each anatomical structure
Hello,

Thanks for this Atlas.

I wonder by any chance there are mask files of each anatomical structure for Waxholm Atlas, better in NifTi format. So i easier apply this atlas in my study. 

Cheers,
Yan
Sep 13, 2022  03:09 PM | Eszter Agnes Papp
RE: mask files of each anatomical structure
Dear Yan,

Thank you for your interest in the Waxholm Space atlas of the Sprague Dawley rat brain.

Do I understand correctly that you need NIfTI files containing one atlas region per file, i.e. separate files for individual regions?

If so, the fastest way to get there would be to use Convert3D, the command-line companion to ITK-SnAP:
http://www.itksnap.org/pmwiki/pmwiki.php?n=Convert3D.Convert3D (download from SourceForge)

Here is how to extract one region (represented by one label):
https://sourceforge.net/p/c3d/git/ci/master/tree/doc/c3d.md#-retain-labels-retain-labels-in-a-label-image

For example, to create a volume containing label 25 only:
c3d atlas.nii -retain-labels 25 -o region25.nii

Best regards,
Eszter
Sep 13, 2022  04:09 PM | ymmpi
RE: mask files of each anatomical structure
Hi Eszter,

Many thanks to your fast reply and introducing this nice tool.
May I ask another question before I start, is Convert3D able to separate each regions to right & left side also?
if not, how should I acheive that?
Best regards,
Yan
Sep 13, 2022  05:09 PM | Eszter Agnes Papp
RE: mask files of each anatomical structure
Hi Yan,

Convert3D does not know, of course, where the left and right side of the brain is situated, but if you identify the voxel coordinates of the sagittal midline, you can define a subvolume to crop the image to using the -region command:
c3d WHS_SD_rat_atlas_v4.nii.gz -region 244x0x0vox 268x1024x512vox -o right-hemisphere.nii.gz

c3d WHS_SD_rat_atlas_v4.nii.gz -region 0x0x0vox 244x1024x512vox -o left-hemisphere.nii.gz

The above coordinates are my best estimate for the midline (see attached image for the cropping results - I tested this on the v3 atlas volume, but it works the same way for version 4, of course).

For details on the -region command, please see https://sourceforge.net/p/c3d/git/ci/master/tree/doc/c3d.md#-region-extract-region-from-image

While looking for this, I found an alternative command for splitting an image volume to individual labels, that may be of interest to you (may be faster if you need all regions):
https://sourceforge.net/p/c3d/git/ci/master/tree/doc/c3d.md#-split-split-multi-label-image-into-binary-images

All the best,
Eszter