Segmenting infant brains across magnetic fields: Domain randomization and annotation curation in ultra-low field MRI

TL;DR

This paper introduces a domain randomization framework and annotation curation techniques to improve brain structure segmentation in ultra-low-field MRI, enabling better neurodevelopmental disorder diagnosis in low-resource settings.

Contribution

It presents a novel domain adaptation method using domain randomization and annotation curation to enhance segmentation accuracy in ultra-low-field MRI.

Findings

Pre-training with domain randomization improves ULF MRI segmentation.

Removing misregistered annotations boosts model performance.

Ensemble predictions achieve competitive segmentation results.

Abstract

Early identification of neurodevelopmental disorders relies on accurate segmentation of brain structures in infancy, a task complicated by rapid brain growth, poor tissue contrast, and motion artifacts in pediatric MRI. These challenges are further exacerbated in ultra-low-field (ULF, 0.064~T) MRI, which, despite its lower image quality, offers an affordable, portable, and sedation-free alternative for use in low-resource settings. In this work, we propose a domain randomization (DR) framework to bridge the domain gap between high-field (HF) and ULF MRI in the context of the hippocampi and basal ganglia segmentation in the LISA challenge. We show that pre-training on whole-brain HF segmentations using DR significantly improves generalization to ULF data, and that careful curation of training labels, by removing misregistered HF-to-ULF annotations from training, further boosts performance. By fusing the predictions of several models through majority voting, we are able to achieve competitive performance. Our results demonstrate that combining robust augmentation with annotation quality control can enable accurate segmentation in ULF data. Our code is available at https://github.com/Medical-Image-Analysis-Laboratory/lisasegm