Abstract:Purpose: Deformable registration problems are conventionally posed in a regularized optimization framework, where balance between fidelity and prescribed regularization usually needs to be tuned for each case. Even so, using a single weight to control regularization strength may be insufficient to reflect spatially variant tissue properties and limit registration performance. In this study, we proposed to incorporate a spatially variant deformation prior into image registration framework using a statistical generative model. Approach: A generator network is trained in an unsupervised setting to maximize the likelihood of observing the moving and fixed image pairs, using an alternating back-propagation approach. The trained model imposes constraints on deformation and serves as an effective low-dimensional deformation parametrization. During registration, optimization is performed over this learned parametrization, eliminating the need for explicit regularization and tuning. The proposed method was tested against SimpleElastix, DIRNet, and Voxelmorph. Results: Experiments with synthetic images and simulated CTs showed that our method yielded registration errors significantly lower than SimpleElastix and DIRNet. Experiments with cardiac magnetic resonance images showed that the method encouraged physical and physiological feasibility of deformation. Evaluation with left ventricle contours showed that our method achieved a dice of (0.93 +/- 0.03) with significant improvement over all SimpleElastix options, DIRNet, and VoxelMorph. Mean average surface distance was on millimeter level, comparable to the best SimpleElastix setting. The average 3D registration time was 12.78 s, faster than 24.70 s in SimpleElastix. Conclusions: The learned implicit parametrization could be an efficacious alternative to regularized B-spline model, more flexible in admitting spatial heterogeneity. (C) 2020 Society of Photo-Optical Instrumentation Engineers (SPIE).

Efficient Laplace Approximation for Bayesian Registration Uncertainty Quantification

Registration Uncertainty Quantification Via Low-Dimensional Characterization of Geometric Deformations.

Frequency Diffeomorphisms for Efficient Image Registration

Quantifying Registration Uncertainty With Sparse Bayesian Modelling

NPBDREG: Uncertainty Assessment in Diffeomorphic Brain MRI Registration using a Non-parametric Bayesian Deep-Learning Based Approach

Data-Driven Model Order Reduction For Diffeomorphic Image Registration

Sliding at first order: Higher-order momentum distributions for discontinuous image registration

Landmark and Intensity Based Registration with Large Deformations via Quasi-conformal Maps

Sliding at First-Order: Higher-Order Momentum Distributions for Discontinuous Image Registration

Bayesian Inference Using the Proximal Mapping: Uncertainty Quantification Under Varying Dimensionality

INR-LDDMM: Fluid-based Medical Image Registration Integrating Implicit Neural Representation and Large Deformation Diffeomorphic Metric Mapping

Bi-level Probabilistic Feature Learning for Deformable Image Registration

Efficient Bayesian Computation by Proximal Markov Chain Monte Carlo: When Langevin Meets Moreau

Iterative Fitting after Elastic Registration: an Efficient Strategy for Accurate Estimation of Parametric Deformations

Bayesian Registration of Functions and Curves

Local Statistical Deformation Models for Deformable Image Registration.

Bayesian Intrinsic Groupwise Image Registration: Unsupervised Disentanglement of Anatomy and Geometry.

Imposing Implicit Feasibility Constraints on Deformable Image Registration Using a Statistical Generative Model.

A Deep Learning Framework for Diffeomorphic Mapping Problems via Quasi-conformal Geometry Applied to Imaging

Distributed-memory large deformation diffeomorphic 3D image registration

Unsupervised Probabilistic Deformation Modeling for Robust Diffeomorphic Registration