Zero-shot Multi-Contrast Brain MRI Registration by Intensity Randomizing T1-weighted MRI (LUMIR25)

TL;DR

This paper introduces a zero-shot brain MRI registration method that generalizes across different contrasts and domain shifts using intensity randomization and other strategies, achieving top performance in the LUMIR25 challenge.

Contribution

The paper proposes a novel zero-shot registration approach that enhances cross-contrast robustness without needing multi-contrast training data, extending prior monomodal registration techniques.

Findings

Achieved 1st place in LUMIR25 challenge.

Significantly improved cross-contrast registration accuracy.

Demonstrated robustness without explicit image synthesis.

Abstract

In this paper, we present our submission to the LUMIR25 task of Learn2Reg 2025, which ranked 1st overall on the test set. Extended from LUMIR24, this year's task focuses on zero-shot registration under domain shifts (e.g., high-field MRI, pathological brains, and various MRI contrasts), while the training data comprises only in-domain T1-weighted brain MRI. We start with a meticulous analysis of LUMIR24 winners to identify the main contributors to strong monomodal registration performance. We highlight the importance of registration-specific inductive biases, including multi-resolution pyramids, inverse and group consistency, topological preservation or diffeomorphism, and correlation-based correspondence establishment. To further generalize to diverse contrasts, we employ three simple but effective strategies: (i) a multimodal loss based on the modality-independent neighborhood descriptor (MIND), (ii) intensity randomization for unseen contrast augmentation, and (iii) lightweight instance-specific optimization (ISO) on feature encoders at inference time. On the validation set, the proposed approach substantially improves T1-T2 registration accuracy, demonstrating robust cross-contrast generalization without relying on explicit image synthesis. These results suggest a practical step toward a registration foundation model that can leverage a single training domain yet remain robust across domain shifts.