CALIBRATION OF THE MAGNETIC TRACKER FOR ACQUISITION OF 3D IMAGES TO GUIDE NEUROSURGERIES

摘要

This work is about the calibration of the magnetic tracker MiniBird™ used to register spatial positions. This calibration is a procedure done necessarily to obtain important points for the execution of Computer Assisted Neuro-Surgery using 3D Images (CANS); as well as the reconstruction of three-dimensional objects inside a Virtual Reality environment. Following the calibration, the movements acquired by the sensor are tracked by the computer in Real time. The inferior tip point of the probe is calculated as the work point. We calculate the calibration matrix as the coordinates of each point corresponding with the coordinates of the magnetic tracker. Obtaining the inferior tip point of the calibration probe requires the calculation of a transformed homogeneous matrix. We researched two minimization methods: (1) The algorithm of Levenberg-Marquardt and (2) The algorithm of Singular Value Decomposition (SVD). The SVD method was chosen because the SVD has a smaller error than Levenberg-Marquardt algorithm when used with a high number of input samples. The calibration software was installed as a plug-in inside the Neuro-Navigator surgical Software.