Unsupervised Learning for Fast Probabilistic Diffeomorphic Registration

TL;DR

This paper introduces an unsupervised probabilistic learning method for fast, diffeomorphic 3D brain registration using CNNs, achieving state-of-the-art accuracy with uncertainty estimation.

Contribution

It presents a novel probabilistic generative model and inference algorithm for unsupervised learning of diffeomorphic registration, combining speed, accuracy, and uncertainty quantification.

Findings

State-of-the-art registration accuracy achieved

Fast runtimes demonstrated

Provides uncertainty estimates and guarantees topology preservation

Abstract

Traditional deformable registration techniques achieve impressive results and offer a rigorous theoretical treatment, but are computationally intensive since they solve an optimization problem for each image pair. Recently, learning-based methods have facilitated fast registration by learning spatial deformation functions. However, these approaches use restricted deformation models, require supervised labels, or do not guarantee a diffeomorphic (topology-preserving) registration. Furthermore, learning-based registration tools have not been derived from a probabilistic framework that can offer uncertainty estimates. In this paper, we present a probabilistic generative model and derive an unsupervised learning-based inference algorithm that makes use of recent developments in convolutional neural networks (CNNs). We demonstrate our method on a 3D brain registration task, and provide an empirical analysis of the algorithm. Our approach results in state of the art accuracy and very fast runtimes, while providing diffeomorphic guarantees and uncertainty estimates. Our implementation is available online at http://voxelmorph.csail.mit.edu .