Low Rank Groupwise Deformations for Motion Tracking in Cardiac Cine MRI

TL;DR

This work introduces a low-rank, groupwise deformation framework for motion tracking in cardiac cine MRI by fusing diffeomorphic image registration with Robust PCA. The method extends pairwise registration to group-wise sequences, producing aligned frames that share a low-rank structure while closely matching a target image, and it employs an over-relaxed ADMM scheme with Sherman–Morrison inverses and Fourier-domain solvers. Empirical results on the ACDC and BioBank datasets show competitive Dice scores and clear low-rank deformations, with group-wise registration outperforming a baseline diffeomorphic approach in both accuracy and efficiency. The study also provides a thorough ablation of key hyperparameters, demonstrating practical guidelines for achieving high-accuracy group-wise motion tracking with controllable rank in the warp fields.

Abstract

Diffeomorphic image registration is a commonly used method to deform one image to resemble another. While warping a single image to another is useful, it can be advantageous to warp multiple images simultaneously, such as in tracking the motion of the heart across a sequence of images. In this paper, our objective is to propose a novel method capable of registering a group or sequence of images to a target image, resulting in registered images that appear identical and therefore have a low rank. Moreover, we aim for these registered images to closely resemble the target image. Through experimental evidence, we will demonstrate our method's superior efficacy in producing low-rank groupwise deformations compared to other state-of-the-art approaches.

Figures (17)

• Figure 1: Diffeomorphic Image Registration of a Circle to a C
• Figure 2: Deformation Field for Circle to C Visualised
• Figure 3: Diffeomorphic Image Registration on Cardiac MRI
• Figure 4: Deformation Field for DiffIR on Patient001
• Figure 5: Diffeomorphic Image Registration on Multiple Cardiac Patients