Semiautomated Registration Using New Markers for Assessing the Accuracy of a Navigation System

Abstract

Objectives: New markers are described that can be used for an improved registration procedure and for the exact comparison of navigation systems. The advantages of the markers are demonstrated, together with an automated segmentation algorithm for locating me centroid of the markers in image space. Compared to manual registration, this method shows an improved registration accuracy.

Materials and Methods: The new markers are detected completely automatically within all scan images. This allows a semiautomatic registration, as a preregistration is performed via the algorithm. Furthermore, the exact coordinates within one scan slice are now determined with the calculation procedure. The calculated data from the preregistration were matched up with a manual preregistration and some reference data, so as to confirm the quality of this new algorithm. The accuracies of several manual and semiautomatic registrations were also compared.

Results: The average deviation between the coordinates of the algorithm and the reference data (coordinate measuring machine) was 0.3 mm. The standard deviation amounted to 0.131 mm. Comparing several manual registrations with the reference data showed that the middle fiducial registration error (FRE) was between 0.7 and 2 mm. In comparison, the FRE remained constant at around 0.7 mm for the semiautomatic registration procedure.

Conclusions: The measured results show a significant improvement in the preregistration data using the new markers together with the algorithm. This improvement leads to a reproducible and more accurate registration. The combination of the new marker type with the automated segmentation algorithm minimizes the human error factor, and provides the opportunity to directly compare image-guided and robotic systems.

Key Words:

• accuracy
• navigation system
• robot
• marker
• automated registration