Good point. In this case the only way I can think of is to create two volumes for each model, and use Segment Registration. You only need the volumes to have a grid that the volume-based distance map registration can use.
I don’t have time now but the first idea that comes to mind is to load a sample dataset, and move it to overlap the models (create transform, set as parent, show volume in volume rendering, show ROI, see if it overlaps). Then harden the transform, clone that volume in Data module, and use those two for Segment Registration.
I’m in the process of implementing an easy way to specify volume geometry from segmentations/models/etc, but it’s not ready yet and will be even more time before we can use it to create empty reference volumes, so for the time being we have to use an existing volume for example as I described above.
Others feel free to chime in.