Futher optimisations
- Store shifted obstructions in a more memory-efficient manner
- Detect whether spins are distant at an earlier stage for annuli
- Do not use ray_grid_intersection if origin and destination are in the same voxel
- Use SVector to tell compiler about number of MRIproperties
- Also, fixes an issue where insideMRI was not respecting the repeats of transformations
Edited by Michiel Cottaar