Propagation of calibration errors in prospective motion correction using external tracking.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Physics

Background:

• Prospective motion correction in MRI, particularly for head imaging, increasingly relies on external tracking devices like cameras.
• Accurate transformation of external tracking data into the MRI scanner's reference frame necessitates precise camera-to-scanner pose estimation (cross-calibration).

Purpose of the Study:

• To investigate the impact of cross-calibration errors on the accuracy of motion correction feedback in MRI.
• To quantify the relationship between calibration inaccuracies and residual motion tracking errors.

Main Methods:

• Derived an operator equation linking calibration errors to motion compensation errors.
• Utilized spherical symmetry and Taylor approximation for small rotations to derive a closed-form expression for residual tracking error.
• Conducted experiments to validate theoretical predictions.

Main Results:

• Confirmed a bilinear dependence of residual rotational error on calibration error and motion magnitude, modulated by the angle between error and motion axes.
• Established bounds for residual translational error based on rotational and translational calibration errors and head displacement.
• Experimental results aligned with theoretical predictions.

Conclusions:

• The study provides a framework for determining necessary cross-calibration accuracy for external tracking devices based on expected motion.
• Less stringent calibration accuracy is acceptable for scans with minimal anticipated movement compared to those with significant motion.
• Clinical applications typically demand calibration accuracy below 1 mm and 1° for effective motion correction.