Dear Yuqian, EasyReg gives you the RAS coordinates directly, rather than the shift in ras (which I believe —inras expects?). You can compute the shift as the warp values minus the ras coordinates at each voxel (pseudocode below) Cheers, /Eugenio

For voxel at I, j, k [r,a,s,1]^t = affine_matrix_header * [I, j, k, 1] warp_new[I, j, k, 0] = r - warp[I,j,k,0] warp_new[I, j, k, 1] = r - warp[I,j,k,1] warp_new[I, j, k, 2] = r - warp[I,j,k,2]

--

Juan Eugenio Iglesias

http://www.jeiglesias.com

From: Chen, Yuqian ychen128@bwh.harvard.edu Date: Wednesday, December 3, 2025 at 23:03 To: freesurfer@nmr.mgh.harvard.edu freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Problem with applying FreeSurfer EasyAtlas warp format

Dear FreeSurfer users,

I’m attempting to use mri_warp_convert to convert an EasyAtlas transform into ITK format, but the resulting transform is incorrect when applied or visualized in 3D Slicer (see image below). The deformation field shrinks the brain, suggesting a possible scaling issue in the converted transform.

Here is the command line I used:

mri_warp_convert --inras sub-05_brain_warp.nii.gz --outitk sub-05_brain_warp-itk.nii.gz -g sub-05_brain.nii.gzhttp://.nii.gz/

[Image]

Has anyone encountered this behavior or knows whether additional steps are needed when converting EasyAtlas transforms to ITK format?

Thank you for any guidance.

Regards,

Yuqian Chen, Ph.D.

Postdoctoral Research Fellow, Harvard Medical School

399 Revolution Drive, Suite 1180, Somerville, MA 02145

ychen128@bwh.harvard.edumailto:ychen128@bwh.harvard.edu