NimbleReg: A light-weight deep-learning framework for diffeomorphic image registration
Authors:
Antoine Legouhy,
Ross Callaghan,
Nolah Mazet,
Vivien Julienne,
Hojjat Azadbakht,
Hui Zhang
Abstract:
This paper presents NimbleReg, a light-weight deep-learning (DL) framework for diffeomorphic image registration leveraging surface representation of multiple segmented anatomical regions. Deep learning has revolutionized image registration but most methods typically rely on cumbersome gridded representations, leading to hardware-intensive models. Reliable fine-grained segmentations, that are now a…
▽ More
This paper presents NimbleReg, a light-weight deep-learning (DL) framework for diffeomorphic image registration leveraging surface representation of multiple segmented anatomical regions. Deep learning has revolutionized image registration but most methods typically rely on cumbersome gridded representations, leading to hardware-intensive models. Reliable fine-grained segmentations, that are now accessible at low cost, are often used to guide the alignment. Light-weight methods representing segmentations in terms of boundary surfaces have been proposed, but they lack mechanism to support the fusion of multiple regional mappings into an overall diffeomorphic transformation. Building on these advances, we propose a DL registration method capable of aligning surfaces from multiple segmented regions to generate an overall diffeomorphic transformation for the whole ambient space. The proposed model is light-weight thanks to a PointNet backbone. Diffeomoprhic properties are guaranteed by taking advantage of the stationary velocity field parametrization of diffeomorphisms. We demonstrate that this approach achieves alignment comparable to state-of-the-art DL-based registration techniques that consume images.
△ Less
Submitted 10 March, 2025;
originally announced March 2025.
