An Eye-Gaze-Controlled Needle Deployment Robot: Design, Modeling, and Experimental Evaluation

Abstract

Robot assisted minimally invasive needle percutaneous intervention promotes the biopsy operation accuracy and speed as well. Yet, Leksell-based needle displacement platform is bulky and unsuitable for certain applications such as MRI guided surgery. Also, controllers such as joysticks or keyboards conventionally used by surgeons to command robotic systems require extra preoperative training and intraoperative sterilization. As a step forward to improve the current robot assisted biopsy system, a compact four degree of freedom needle deployment robot is designed, and a deep learning-based eye gaze control method, using only one camera, is derived to facilitate the interaction efficiency and efficacy. A first-generation prototype is constructed to evaluate the performance of the proposed eye gaze controlled needle deployment method. Experimental results demonstrate that the proposed system can achieve satisfactory positioning precision, with a maximum average error of 0.3883°, and it is capable of tracking the eye fixation angle in real-time.