Related Experiment Video
Updated: Jan 17, 2026
06:56
Human Fetal Blood Flow Quantification with Magnetic Resonance Imaging and Motion Compensation
Published on: January 7, 2021
2.8K
An Integrated Automatic Framework for Super-Resolution Reconstruction of Motion-Corrupted Fetal Brain MRI With Prior
IEEE Transactions on Bio-Medical Engineering
|September 17, 2025
Summary
This study introduces an automated method for high-resolution fetal brain MRI reconstruction, improving prenatal diagnostics. The novel framework utilizes anatomical priors for enhanced accuracy in super-resolution reconstruction (SRR).
Area of Science:
- Medical Imaging
- Neuroimaging
- Computational Anatomy
Background:
- Super-resolution reconstruction (SRR) of fetal brain MRI is crucial for prenatal diagnostics and developmental assessment.
- Conventional SRR methods require manual intervention and struggle with motion-corrupted, thick-slice data, leading to blurred images.
- Existing techniques are limited by challenges in extracting cerebral structures and handling indistinct tissue boundaries in fetal MRI.
Purpose of the Study:
- To develop an automated fetal brain SRR framework that integrates individual and group-level priors.
- To enhance the accuracy and quality of super-resolution reconstructed fetal brain MRI.
- To overcome limitations of conventional methods in handling motion artifacts and indistinct boundaries.
Main Methods:
- Developed a robust fetal brain extraction method using the Segment Anything Model (SAM).
- Integrated tissue segmentation-derived anatomical priors into slice-to-volume registration and volumetric reconstruction.
- Employed an iterative optimization scheme alternating between registration and reconstruction, guided by longitudinal fetal brain atlases.
Main Results:
- The proposed framework demonstrated significantly improved quantitative and qualitative performance in fetal brain SRR.
- Integration of anatomical priors enhanced boundary alignment during registration and mitigated misalignment issues.
- Utilizing longitudinal atlases enriched structural details and prevented reconstruction of outliers in the MR images.
Conclusions:
- The developed automatic SRR framework effectively addresses challenges in fetal brain MRI reconstruction.
- The integration of individual and group-level priors significantly improves the quality and accuracy of reconstructed images.
- This approach holds promise for more comprehensive prenatal examinations and precise fetal brain development quantification.
