• Corpus ID: 239885402

Uncertainty quantification in non-rigid image registration via stochastic gradient Markov chain Monte Carlo

  title={Uncertainty quantification in non-rigid image registration via stochastic gradient Markov chain Monte Carlo},
  author={Daniel Grzech and Mohammad Farid Azampour and Huaqi Qiu and Ben Glocker and Bernhard Kainz and Lo{\"i}c Le Folgoc},
We develop a new Bayesian model for non-rigid registration of three-dimensional medical images, with a focus on uncertainty quantification. Probabilistic registration of large images with calibrated uncertainty estimates is difficult for both computational and modelling reasons. To address the computational issues, we explore connections between the Markov chain Monte Carlo by backpropagation and the variational inference by backpropagation frameworks, in order to efficiently draw samples from… 

Figures and Tables from this paper


Bayesian Characterization of Uncertainty in Multi-modal Image Registration
A general formalism is proposed to quantify Bayesian uncertainty in the registration of multi-modal images through an extended probability model that introduces and then marginalizes out a stochastic transfer function between moving and fixed image intensities.
A multilevel Markov Chain Monte Carlo approach for uncertainty quantification in deformable registration
This work estimates the registration uncertainty based on Bayesian analysis of the MCMC approach into a cost reducing multilevel framework, and demonstrates its correctness by comparison with a ground-truth of the posterior distribution, and shows the reliability as uncertainty estimator on brain MRI images.
Probabilistic Diffeomorphic Registration: Representing Uncertainty
A novel mathematical framework for representing uncertainty in large deformation diffeomorphic image registration by estimating the full posterior distribution in order to represent uncertainty, in contrast to methods in which the posterior is approximated via Monte Carlo sampling or maximized in maximum a-posteriori (MAP) estimation.
Summarizing and Visualizing Uncertainty in Non-rigid Registration
This paper shows how to determine the registration uncertainty, as well as the most likely deformation, by using an elastic Bayesian registration framework that generates a dense posterior distribution on deformations, and introduces methods that summarize the high-dimensional uncertainty information.
Bayesian Estimation of Regularization and Atlas Building in Diffeomorphic Image Registration
An atlas estimation procedure that simultaneously estimates the parameters controlling the smoothness of the diffeomorphic transformations and a Monte Carlo Expectation Maximization algorithm, where the expectation step is approximated via Hamiltonian Monte Carlo sampling on the manifold of diffeomorphisms.
Unsupervised Deep-Learning Based Deformable Image Registration: A Bayesian Framework
This work introduces a fully Bayesian framework for unsupervised DL-based deformable image registration and provides an estimate of the uncertainty in the deformation-field by characterizing the true posterior distribution, thus, avoiding potential over-fitting.
Probabilistic Image Registration via Deep Multi-class Classification: Characterizing Uncertainty
This work presents a novel approach to probabilistic image registration that leverages the strengths of deep-learning for modeling agreement between images, and enables local predictions of registration uncertainty and diagnostics that can indicate areas that seem unrelated in the two images.
Learning a Probabilistic Model for Diffeomorphic Registration
The unsupervised method uses a conditional variational autoencoder network and constrain transformations to be symmetric and diffeomorphic by applying a differentiable exponentiation layer with a symmetric loss function and provides multi-scale velocity field estimations.
Metric Learning for Image Registration
This work embeds a deep learning model in an optimization-based registration algorithm to parameterize and data-adapt the registration model itself, allowing controlling the desired level of regularity and preserving structural properties of a registration model.
A Markov Chain Monte Carlo based rigid image registration method
This work formulate the image registration problem within a Bayesian framework and generate samples from the resulting posterior density of the registration parameters using MCMC, and posterior density is characterized through the samples that are drawn with the MCMC principle.