[Neuroimaging] non linear deformation

[Ariel Rokem]

Hi Romain,

Dipy has a full implementation of the symmetric diffeomorphic registration:
http://nipy.org/dipy/examples_built/syn_registration_3d.html

If you already have your deformation field, you can use that to initialize
a dipy.align.imwarp.DiffeomorphicMap object and apply the transformation to
your input image.

Hope that helps,

Ariel

On Wed, Sep 19, 2018 at 6:51 AM Alexis Roche <alexis.roche at gmail.com> wrote:

> Hi Romain,
>
> If it’s just to apply a deformation field, one easy way to go is to use
> scipy.ndimage.map_coordinates
>
> Cheers
>
> On Wed 19 Sep 2018 at 14:26, valabregue <romain.valabregue at upmc.fr> wrote:
>
>> Dear all
>>
>> Can anyone point me out some python library to apply non linear
>> deformation to the image ?
>>
>> This is for data augmentation so I do not need to do a non linear
>> registration, but just a way to apply a random smooth deformation field.
>>
>> Is there in nibabel support for reading deformation field (from spm or
>> ants) and then to apply it ?
>>
>>
>> Many thanks
>>
>> Romain
>>
>> _______________________________________________
>> Neuroimaging mailing list
>> Neuroimaging at python.org
>> https://mail.python.org/mailman/listinfo/neuroimaging
>>
> --
> https://sites.google.com/site/alexisroche
> _______________________________________________
> Neuroimaging mailing list
> Neuroimaging at python.org
> https://mail.python.org/mailman/listinfo/neuroimaging
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/neuroimaging/attachments/20180922/f4013cc3/attachment.html>