Autofocusing+: Noise-Resilient Motion Correction in Magnetic Resonance Imaging

TL;DR

Autofocusing+ is a neural network-enhanced method that improves motion artifact correction in MRI by combining optimization routines with deep learning priors, demonstrating robustness to noise and anatomical variations.

Contribution

We introduce a neural network-based regularization for Autofocusing, enhancing its noise resilience and convergence speed in motion artifact removal for MRI.

Findings

Outperforms state-of-the-art demotion methods

Resilient to noise and anatomical variations

Validated on synthetic and real noisy data

Abstract

Image corruption by motion artifacts is an ingrained problem in Magnetic Resonance Imaging (MRI). In this work, we propose a neural network-based regularization term to enhance Autofocusing, a classic optimization-based method to remove motion artifacts. The method takes the best of both worlds: the optimization-based routine iteratively executes the blind demotion and deep learning-based prior penalizes for unrealistic restorations and speeds up the convergence. We validate the method on three models of motion trajectories, using synthetic and real noisy data. The method proves resilient to noise and anatomic structure variation, outperforming the state-of-the-art demotion methods.