There are many preprocessing tasks that could be easily implemented with the core functionality available in nitorch. For example:
-Gaussian smoothing
-Resampling of image data (i.e. change the voxel size)
-Reslice to another image's grid
-Cropping
How should this be implemented? I suggest just implementing the core functions operating on an image to image basis and let the user code up their own multi-subject processing pipelines.
There are many preprocessing tasks that could be easily implemented with the core functionality available in nitorch. For example:
-Gaussian smoothing -Resampling of image data (i.e. change the voxel size) -Reslice to another image's grid -Cropping
How should this be implemented? I suggest just implementing the core functions operating on an image to image basis and let the user code up their own multi-subject processing pipelines.