Set-Based Groupwise Registration for Variable-Length, Variable-Contrast Cardiac MRI

2026-05-11

AI Summary

Existing deep learning–based groupwise registration methods are constrained by fixed input architectures, limiting their generalization to cardiac MRI sequences with varying lengths, contrasts, and temporal orders. To address this, this work proposes AnyTwoReg, a novel framework that, for the first time, models MRI sequences as unordered sets. By employing a shared encoder and correlation-guided feature aggregation, AnyTwoReg constructs a permutation-invariant reference while learning a permutation-equivariant deformation field. This design decouples the registration process from dependencies on sequence length and input ordering. Furthermore, by leveraging a pretrained foundation model to extract contrast-robust features, the method achieves zero-shot generalization across imaging protocols. Trained solely on a single T₁-weighted dataset, AnyTwoReg effectively registers diverse quantitative MRI sequences—such as MOLLI and ASL—with varying frame counts (11–60) and contrast properties, substantially improving downstream quantitative imaging quality.

Abstract

Quantitative cardiac magnetic resonance imaging (MRI) enables non-invasive myocardial tissue characterization but relies on robust motion correction within these variable-length, variable-contrast image sequences. Groupwise registration, which simultaneously aligns all images, has shown greater robustness than pairwise registration for motion correction. However, current deep-learning-based groupwise registration methods cannot generalize across MRI sequences: the architecture typically encodes input data as a fixed-length channel stack, which rigidly couples network design to protocol-specific sequence length, input ordering, and contrast dynamics. At inference time, any change in imaging protocols will render the network unusable. In this work, we introduce \emph{\AnyTwoReg}, a new set-based groupwise registration framework that takes a quantitative MRI sequence as an unordered set. This set formulation fundamentally decouples network design from sequence length and input ordering. By utilizing a shared encoder and correlation-guided feature aggregation, \emph{\AnyTwoReg} constructs a permutation-invariant canonical reference for registration, and learns a permutation-equivariant mapping from images to deformation fields. Additionally, we extract contrast-insensitive image features from an existing foundation model to handle extreme contrast variations. Trained exclusively on a single public $T_1$ mapping dataset (STONE, sequence length $L=11$), \AnyTwoReg generalizes to two unseen quantitative MRI datasets (MOLLI, ASL) with variable lengths ($L \in [11, 60]$) and different contrast dynamics. It achieves strong cross-protocol generalization in a zero-shot manner, and consistently improves downstream quantitative mapping quality. Notably, while designed for quantitative MRI sequences, our framework is directly applicable to Cine MRI sequences for inter-cardiac-phase registration.

Problem

• groupwise registration
• cardiac MRI
• variable-length sequences
• contrast variation
• cross-protocol generalization

Innovation

• set-based registration
• groupwise registration
• permutation invariance
• cross-protocol generalization
• contrast-invariant features