[bokkyuki@usc.edu]

Hello, thank you for your reply. I am trying to identify stroke lesion. So I want to make a label of stroke lesion on each individual, and transform the lesion label to standard atlas space to compare the lesions among different stroke patients.

[Anand Joshi]

I am not sure what can be done by putting lesions into common space? If you want to perform measurements of lesions, such as size etc, then you can do that in the original space itself.

If you really have to transform the lesion labels to a common space, then you can do that using the inverse map stored in SVReg directory (subbasename.svreg.inv.map.nii.gz).

You can use matlab with the utility functions available from

http://neuroimage.usc.edu/neuro/Resources/BST_SVReg_Utilities
assuming your subbasename is name of the subject prefix, including path.

e.g. if your subject file is c:\Users\ajoshi\Downloads\sub01\sub01.nii.gz then subbasename would be c:\Users\ajoshi\Downloads\sub01\sub01

and

atlasbasename=C:\Program Files\BrainSuite15b\svreg\BrainSuiteAtlas1\mri

The following matlab code should do the warping.
vmap=load_nii_z(‘subbasename.svreg.inv.map.nii.gz’);
vw=load_nii_z(‘atlasbasename.label.nii.gz’);
vlab=load_nii_z(‘subbasename.label.nii.gz’)

vw.img=interp3(vlab.img,vmap.img(:,:,:,2),vmap.img(:,:,:,1),vmap.img(:,:,:,3),’nearest’);

save_nii(vlab,’warped_labels.nii.gz’);

[bokkyuki@usc.edu]

Hi,

Thank you so much for your support. The reason why I am trying to convert the individual image to standard space is to identify the lesion location, and compare the location among participants.

Anyway, the code is working very well.

[Anand Joshi]

Hi, Can you please explain more what you are trying to do? Are you trying to transform the individual anatomy to common atlas space? SVReg performs registration of subject and atlas and warps atlas to match the subject. The labels from warped atlas are then transferred to the subject anatomy. Can you please explain more about what you are trying to do?

• This reply was modified 10 years ago by Anand Joshi.

[Author]

Posts