search

TIB-Lab / MouseX-Allen-Atlas

Sex matters: the MouseX DW-ALLEN Atlas and Parcellation for mice diffusion-weighted MR imaging
GNU General Public License v3.0
0 stars 0 forks source link

Voxel resolution/affine scaling #1

Open araikes opened 6 months ago

araikes commented 6 months ago

Good afternoon,

I'm very interested in being able to use this template/atlas in some of my research. I've been using ANTs to do the registration between my images and your template and generate consistent failures (though this is not an ANTs question).

As I've been examining your image, I note that your voxel resolution for the T2w template is 0.6x0.7x4 on a 258x240x14 grid. Generating a mask from that image yields a brain volume of 572,200 mm3 (comparable to human brain volumes).

Looking at your preprint, these resolution features don't match so I'm curious if there was some behind-the-scenes affine scaling that isn't being reflected in the actual NIFTI header yield such large volumes and and the inability to accurately register.

Thanks

TIB-Lab commented 4 months ago

Good afternoon!

First of all please accept our apologies for not answering earlier, we are new to GitHub and haven't noticed :( But we are so happy you find this template/atlas useful for your experiments!

Now, concerning your question, you are right, we did rescale. The pixel dimension is 0.069x0.07x0.4 mm = 0.001932 mm3. Whole brain volume is thus 490.32 mm3.

Concerning your problems using ANTs, don't hesitate write us to see if we can somehow help! We are still figuring out how to finish the configurations on GitHub, so you can write me to my personal email: patricia.martinezt@umh.es

Thanks for your interest!

You can check our latest modifications here on the repository, and hopefully soon the published final manuscript. We have reduced the atlas parcellation from 63 to 62 regions due to volume size of one small region, and all the atlases we provide are now corresponding to this corrected parcellation, you may rather use this newest version.