Stop moving: MR motion correction as an opportunity for artificial intelligence

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Biomedical Engineering

Background:

• Subject motion is a major challenge in magnetic resonance imaging (MRI), degrading image quality.
• Numerous prospective and retrospective MRI motion correction techniques exist.
• Deep learning methods have emerged as state-of-the-art for MRI motion correction.

Purpose of the Study:

• To provide a comprehensive review of deep learning-based MRI motion correction.
• To detail neural networks used for artifact reduction and motion estimation.
• To explore the application of motion estimation in downstream tasks.

Main Methods:

• Review of deep learning architectures for MRI motion correction.
• Analysis of neural networks in image and frequency domains.
• Discussion of motion estimation's role in MRI reconstruction and other applications.

Main Results:

• Deep learning approaches demonstrate superior performance in MRI motion correction.
• Various neural network strategies effectively reduce motion artifacts and estimate motion parameters.
• Integration of motion estimation enhances downstream applications beyond reconstruction.

Conclusions:

• Deep learning offers powerful solutions for MRI motion correction challenges.
• Further research is needed to address current limitations and explore future directions.
• Enhanced interaction between research areas can advance MRI technology.