Output details

Probabilistic inference of regularisation in non-rigid registration

This work forms the first mathematically principled approach in the literature to infer data-driven regularization of nonlinear registration using a Variational Bayesian framework. It originally was accepted for oral presentation (3% acceptance rate) at MICCAI'11, an internationally well-established, double-blind peer-reviewed conference in the field, and ultimately published in a leading neuroimaging journal. This work is already impacting the derivation of other registration properties used for stabilizing and steering image registration technology in general, such as by researchers at Harvard University (details available).