Multi-task Magnetic Resonance Imaging Reconstruction using Meta-learning

TL;DR

This paper introduces a meta-learning approach for MRI reconstruction that effectively handles multiple imaging sequences, improving the reconstruction quality of highly-undersampled data across diverse datasets.

Contribution

The proposed method is the first to apply meta-learning for multi-task MRI reconstruction, enabling simultaneous learning from multiple datasets with different contrasts.

Findings

Outperforms existing single-task reconstruction methods.

Successfully reconstructs highly-undersampled MRI data from various datasets.

Demonstrates improved generalization across different imaging sequences.

Abstract

Using single-task deep learning methods to reconstruct Magnetic Resonance Imaging (MRI) data acquired with different imaging sequences is inherently challenging. The trained deep learning model typically lacks generalizability, and the dissimilarity among image datasets with different types of contrast leads to suboptimal learning performance. This paper proposes a meta-learning approach to efficiently learn image features from multiple MR image datasets. Our algorithm can perform multi-task learning to simultaneously reconstruct MR images acquired using different imaging sequences with different image contrasts. The experiment results demonstrate the ability of our new meta-learning reconstruction method to successfully reconstruct highly-undersampled k-space data from multiple MRI datasets simultaneously, outperforming other compelling reconstruction methods previously developed for single-task learning.