Pancakes: Consistent Multi-Protocol Image Segmentation Across Biomedical Domains

John Guttag (Massachusetts Institute of Technology) · Marianne Rakic (CSAIL, MIT) · Siyu Gai (Massachusetts Institute of Technology) · Etienne Chollet (Massachusetts Institute of Technology) · Adrian Dalca (MIT, HMS)
anatomical regionsautomatic segmentation modelsbiomedical image segmentationbrain mricoherent segmentationdiverse segmentation mapslabor-intensive promptingmulti-label segmentationsegmentation protocolssemantic consistencystrong baselinestissue typesunseen domainsvascular territories

A single biomedical image can be segmented in multiple valid ways, depending on the application. For instance, a brain MRI may be divided according to tissue types, vascular territories, broad anatomical regions, fine-grained anatomy, or pathology. Existing automatic segmentation models typically either (1) support only a single protocol