Development of Deep Learning Algorithm for Autonomous Vascular Intervention Robot System

Abstract

Manual vascular intervention (VI) procedure has been performed under radiation exposure. To overcome this issue, recently many commercial master-slave VI robotic systems have been developed. However, master-slave VI robot system still have issues to resolve. The operator must reside near the master device and control the slave robot only through the master device. In addition, the operator must simultaneously recognize the surgical tool from the X-ray image along with the operation of the master device. To solve the limitations of master-slave VI robot system, we propose an autonomous VI robot system with deep learning algorithm. The proposed autonomous VI robot with deep learning algorithm drives surgical tool to the target blood vessel location while simultaneously performing surgical tool recognition. To verify the effectiveness of the developed autonomous VI robot system, an experiment was conducted using a vascular phantom.