Trained models used in the publication for brain and liver registration.
The liver models correspond to those finetuned in one step, since these yielded better results than the models finetuned in two steps (see publication for more details).

And the SpatialTransformer model needed for the HF demo.

Compiled model with the SpatialTransformer layer, for the HF model

Test data for brain registration. Samples taken for the IXI dataset
Fixed images: ixi_image.nii.gz and ixi_segmentation.nii.gz
Moving images: ixi_image2.nii.gz and ixi_segmentation2.nii.gz