@mattcieslak told me about these transforms that can apparently calculate a map of distances from an ROI very rapidly (e.g., using implementations in simpleitk, ANTS, but also implemented in scipy: https://docs.scipy.org/doc/scipy/reference/generated/scipy.ndimage.distance_transform_edt.html#scipy.ndimage.distance_transform_edt). My thinking is that if this really is very fast, it can be used to calculate this transform once upfront for each of our waypoint/endpoint ROI and potentially would make calculations of distances to streamlines more efficient.
See https://arxiv.org/pdf/2106.03503
@mattcieslak told me about these transforms that can apparently calculate a map of distances from an ROI very rapidly (e.g., using implementations in simpleitk, ANTS, but also implemented in scipy: https://docs.scipy.org/doc/scipy/reference/generated/scipy.ndimage.distance_transform_edt.html#scipy.ndimage.distance_transform_edt). My thinking is that if this really is very fast, it can be used to calculate this transform once upfront for each of our waypoint/endpoint ROI and potentially would make calculations of distances to streamlines more efficient.