No-attachment head tracking method of robotic transcranial magnetic stimulation based on vision sensor

Abstract

A vision-based head tracking method for robotic transcranial magnetic stimulation (TMS), requiring no attachments to patient, is proposed to track the movements of patient's head during treatment. First, a facial landmark detection method based on Constrained Local Neural Field (CLNF) algorithm is used. Second, a method of head location and pose estimation is presented and a filter for image restoration is used for a higher accuracy. Finally, calibration approaches are proposed based on the coordinates of the robot arm and camera and real-time motion tracking system. Experiment results indicate that the accuracy and real-time performance of this head tracking method satisfy the requirements of TMS treatment.