Intersection-based slice motion estimation for fetal brain imaging

Area of Science:

• Medical Imaging
• Radiology
• Biomedical Engineering

Background:

• Fetal MRI is crucial for studying fetal development and early diagnosis.
• Motion artifacts from maternal and fetal movement degrade fetal MRI quality.
• Current methods often involve complex registration and reconstruction, leading to data loss.

Purpose of the Study:

• To develop a novel, reconstruction-independent method for correcting inter-slice motion artifacts in fetal MRI.
• To improve the quality of 3D fetal MRI volumes by addressing motion-induced artifacts.
• To offer an alternative to existing motion correction techniques that may reduce data quality.

Main Methods:

• Utilized the intersections of orthogonal MRI slices to estimate motion.
• Developed a machine learning classifier to identify misaligned slices.
• Employed a multistart optimization approach for parameter correction of misaligned slices.

Main Results:

• Demonstrated very low registration errors on simulated datasets.
• Qualitative analysis on real fetal MRI data showed superior performance compared to existing methods.
• The proposed method effectively corrects motion artifacts without compromising data integrity.

Conclusions:

• The novel reconstruction-independent method significantly improves fetal MRI quality by correcting motion artifacts.
• This approach offers a more robust and effective solution for motion correction in fetal imaging.
• The machine learning-based slice identification enhances the accuracy and reliability of the correction process.