Robust myocardial motion tracking for echocardiography: variational framework integrating local-to-global deformation

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Cardiovascular Ultrasound

Background:

• Accurate left ventricular (LV) border tracking is crucial for echocardiography analysis.
• Conventional optical flow methods struggle with ultrasound artifacts and data corruption, leading to distorted LV contours.
• Existing techniques are susceptible to errors from shadowing and dropouts in cardiac wall imaging.

Purpose of the Study:

• To develop a robust real-time algorithm for myocardial border tracking in echocardiography.
• To address and mitigate shape distortion issues caused by corrupted echocardiographic data.
• To enhance the accuracy and reliability of LV contour tracking in challenging clinical scenarios.

Main Methods:

• Integration of local optical flow motion with global deformation within a variational framework.
• A novel descent method guides tracking points using local speckle patterns and global affine transformation constraints.
• Real-time processing capability for immediate clinical application.

Main Results:

• The proposed algorithm demonstrates superior performance compared to conventional methods in real-world experiments.
• Effective handling of local data corruption and ultrasound artifacts, preventing tracking errors.
• Maintained accurate LV contours even with significant image quality degradation.

Conclusions:

• The developed variational framework offers a robust solution for myocardial border tracking in echocardiography.
• This method significantly improves the accuracy of LV contour delineation, especially under adverse imaging conditions.
• Enhanced tracking accuracy has the potential to improve the reliability of echocardiographic diagnoses.