Autofocusing+: Noise-Resilient Motion Correction in Magnetic Resonance Imaging
This work addresses motion artifacts in MRI, a critical problem for medical imaging, but it is incremental as it builds on existing Autofocusing techniques.
The authors tackled motion artifact removal in MRI by combining a neural network-based regularization term with the classic Autofocusing optimization method, resulting in a noise-resilient approach that outperformed state-of-the-art demotion methods.
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.