Automatic targeting method and accuracy study in robot assisted needle procedures

We present an automatic needle targeting method for fluoroscopy-guided percutaneous access procedures. The approach is derived from the manual needle superimposition technique making it intuitive and familiar to surgeons and radiologists. The proposed algorithm is insensitive to image distortion and does not require any C-Arm calibration or initial pose estimation. Needle alignment is performed using a direct adaptive visual servoing approach; once the desired orientation is achieved, insertion is performed using under joystick control. The algorithm was implemented and tested using our AcuBot robot, purposely built for percutaneous image guided interventions. A series of tests were performed showing that the proposed approach increases accuracy and reduces radiation exposure.