Abstract:The stationary velocity field (SVF) approach allows to build parametrizations of invertible deformation fields, which is often a desirable property in image registration. Its expressiveness is particularly attractive when used as a block following a machine learning-inspired network. However, it can struggle with large deformations. We extend the SVF approach to matrix groups, in particular $\SE(3)$. This moves Euclidean transformations into the low-frequency part, towards which network architectures are often naturally biased, so that larger motions can be recovered more easily. This requires an extension of the flow equation, for which we provide sufficient conditions for existence. We further prove a decomposition condition that allows us to apply a scaling-and-squaring approach for efficient numerical integration of the flow equation. We numerically validate the approach on inter-patient registration of 3D MRI images of the human brain.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in the image registration task, the existing methods based on Stationary Velocity Field (SVF) perform poorly when dealing with large deformations. Specifically: 1. **Limitations of Existing Methods**: Traditional SVF methods perform well when dealing with small deformations, but when dealing with large deformations that include large rotations or translations, the generated deformation fields often deviate from the real situation, resulting in inaccurate registration results. 2. **Nature of the Problem**: This limitation stems from the complexity and non - intuitiveness of SVF methods when representing common affine transformations (such as rotation and translation). These transformations require the network to generate the corresponding vector fields from scratch, increasing the learning difficulty. 3. **Proposed Solution**: To solve this problem, the author proposes to extend the SVF method to matrix groups (especially the SE(3) group). By introducing Euclidean transformations (such as rotation and translation) into the low - frequency part, the network architecture can more easily capture and process larger motions. This not only improves the model's ability to handle large deformations but also retains the advantage of the SVF method in generating invertible deformation fields. ### Specific Improvements - **Stationary Velocity Field on Matrix Groups**: The author introduces a matrix - group - valued velocity field, allowing the use of constant matrix fields to represent common linear transformations (such as rotation), thus simplifying the learning process. - **Extension of the Flow Equation**: To adapt to the matrix - group - valued velocity field, the author extends the flow equation and provides sufficient conditions for existence. In addition, they also prove a decomposition condition, making it possible to perform numerical integration efficiently by the scaling - and - squaring method. - **Numerical Verification**: The author conducts numerical verification on the cross - patient registration task of 3D MRI brain maps, demonstrating the effectiveness of the new method in dealing with large deformations. ### Summary The main goal of this paper is to improve the existing SVF method by introducing a matrix - group - valued velocity field, enabling it to better handle large deformations while maintaining the ability to generate invertible deformation fields. This improvement is expected to improve the accuracy and robustness in image registration tasks, especially in applications such as medical image analysis.

Stationary Velocity Fields on Matrix Groups for Deformable Image Registration

Networks for Joint Affine and Non-parametric Image Registration

A method for large diffeomorphic registration via broken geodesics

Incompressible image registration using divergence-conforming B-splines

Distributed-memory large deformation diffeomorphic 3D image registration

Diffeomorphic Image Registration with Neural Velocity Field

Deformable templates using large deformation kinematics

An Inexact Newton-Krylov Algorithm for Constrained Diffeomorphic Image Registration

Deformable Slice-to-Volume Registration for Motion Correction in Fetal Body MRI

Medical Image Registration using optimal control of a linear hyperbolic transport equation with a DG discretization

CLAIRE: Scalable GPU-Accelerated Algorithms for Diffeomorphic Image Registration in 3D

Non-rigid Registration Algorithm Based on Modally Controlled Free Form Deformation

Joint Deformable Registration of Large EM Image Volumes: A Matrix Solver Approach

Learning Diffeomorphism for Image Registration with Time-Continuous Networks using Semigroup Regularization

Efficient and Robust Registration on the 3D Special Euclidean Group

A Nonrigid Registration Framework Using Spatially Encoded Mutual Information and Free-Form Deformations

Low Rank Groupwise Deformations for Motion Tracking in Cardiac Cine MRI

Viscous Fluid Model-Based Non-Rigid Registration Incorporating Scale Space Keypoints

Deformable Image Registration Using Vision Transformers for Cardiac Motion Estimation from Cine Cardiac MRI Images

Indirect Image Registration with Large Diffeomorphic Deformations