Apply non-random affine

[Constantin-Jehn]

Hello,

what would you suggest as simplest way to apply a specific affine to an image.
That would be the deterministic pendent of RandomAffine.

Thanks a lot.

[fepegar]

Hi, @Constantin-Jehn. You can use Affine for that:

import torchio as tio
colin = tio.datasets.Colin27().t1
colin.plot()

look_up = tio.Affine(1, (45, 0, 0), 0)
colin_looking_up = look_up(colin)
colin_looking_up.plot()

[Constantin-Jehn]

Hi,
thanks a lot for your answer!

I would suggest to add the transform to torchio's documentation, at least I would have expected to find it either under Transforms/PreProcessing/Spatial or Transforms/Others.

Is there also a was to provide the affine-matrix itself as input?

Cheers

[fepegar]

Hi, @Constantin-Jehn. Sorry for the delayed reply. There is currently no way to pass an affine matrix directly.

I think it would be nice to mention in the docs that there are non-random versions of all transforms. Where do you think would be a good place for that?