uniGradICON: A Foundation Model for Medical Image Registration

arXiv:2403.05780v1 Announce Type: new
Abstract: Conventional medical image registration approaches directly optimize over the parameters of a transformation model. These approaches have been highly successful and are used generically for registrations of different anatomical regions. Recent deep registration networks are incredibly fast and accurate but are only trained for specific tasks. Hence, they are no longer generic registration approaches. We therefore propose uniGradICON, a first step toward a foundation model for registration providing 1) great performance \emph{across} multiple datasets which …

arxiv cs.cv foundation foundation model image medical registration type