Evaluation of an image-based tracking workflow with Kalman filtering for automatic image plane alignment in...

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Signal Processing

Background:

• Real-time magnetic resonance (MR) imaging requires precise tracking for image plane alignment.
• Existing workflows utilize tracking devices with micro-coils and passive markers for real-time MR image guidance.
• Kalman filters are effective estimators and predictors suitable for tracking applications.

Purpose of the Study:

• To integrate a Kalman filter into a previously developed MR image plane alignment workflow.
• To enhance the tracking performance and prediction accuracy of the MR tracking device.
• To dynamically adjust Kalman filter parameters based on available 3D pose information.

Main Methods:

• Integration of a Kalman filter into the existing MR image plane alignment workflow.
• Development of a tracking device with 2 resonant micro-coils and a passive marker.
• Dynamic adjustment of Kalman filter measurement noise covariances based on image plane orientation.
• Simulation studies to quantify tracking performance improvements.

Main Results:

• The Kalman filter integration improved the prediction of the tracking device's position and orientation.
• Dynamic adjustment of noise covariances enhanced tracking performance, especially when only partial 3D pose data was available.
• Simulation results demonstrated quantifiable improvements in tracking accuracy.

Conclusions:

• The enhanced Kalman filter workflow shows improved performance for real-time MR image tracking.
• Dynamic covariance adjustment is crucial for accurate pose estimation in MR image-guided interventions.
• Further in-scanner experiments are needed to validate these findings in a clinical setting.