Description
Is your feature request related to a problem? Please describe.
Currently, for image registration task, MONAI only has BendingEnergyLoss to enforce smooth deformations, which is based on the second partials of the displacement fields. A common regularizer used in unpaired brain MR images registration (e.g., VoxelMorph and TransMorph) is the diffusion regularizer, which is based on the first partials of the displacement fields. It would be great if diffusion loss can be included as part of MONAI.
Describe the solution you'd like
Should be straight forward to implement using the existing monai.losses.deform.spatial_gradient.
Describe alternatives you've considered
One could alternatively customize a diffusion loss using monai.losses.deform.spatial_gradient in the training script.
Additional context
Feel free to assign me to this issue. :)