Related Experiment Video
Updated: Jun 28, 2026
06:56
Human Fetal Blood Flow Quantification with Magnetic Resonance Imaging and Motion Compensation
Published on: January 7, 2021
Fully automated motion correction in first-pass myocardial perfusion MR image sequences
Julien Milles1, Rob J van der Geest, Michael Jerosch-Herold
1Division of Image Processing, Department of Radiology, Leiden University Medical Center, 2300 RC Leiden, The Netherlands.
IEEE Transactions on Medical Imaging
|October 29, 2008
Summary
This study introduces an automatic cardiac perfusion MRI registration method using independent component analysis (ICA). The novel approach enhances image accuracy and robustness for clinical applications.
Area of Science:
- Medical Imaging
- Cardiovascular Imaging
- Image Registration
Background:
- Cardiac perfusion MRI is crucial for diagnosing heart conditions.
- Accurate registration of perfusion data is essential for reliable analysis.
- Existing registration methods may lack automation and robustness.
Purpose of the Study:
- To develop and validate a fully automatic registration method for cardiac perfusion MRI.
- To leverage independent component analysis (ICA) for enhanced feature extraction and registration accuracy.
- To assess the method's performance in terms of motion reduction, time-intensity curve accuracy, and clinical parameter agreement.
Main Methods:
- A novel registration method employing ICA to extract physiological features and time-intensity dynamics.
- Computation of a time-varying reference image based on ICA results.
- A two-pass registration framework utilizing the computed reference image.
- Validation using 46 clinical perfusion MRI datasets with diverse motion patterns.
Main Results:
- Significant reduction in average left ventricle (LV) motion from 1.26 pixels to 0.64 pixels.
- Improved time-intensity curve accuracy, with average error reduced from 2.65% to 0.87%.
- Good agreement between clinically relevant parameters derived from registered data and manual gold standards.
- Demonstrated robustness against variations in breathing protocols in simulated tests.
Conclusions:
- The fully automatic ICA-based registration method is accurate and robust for cardiac perfusion MRI.
- The method's speed and reliability are suitable for clinical environments.
- This technique offers potential for improved diagnostic accuracy in cardiovascular imaging.
