Real Time Detection and Tracking of Guide Wire/Catheter for Interventional Embolization Robot based on Deep Learning

Interventional embolization of tumors is currently a widely used treatment method for patients. Interventional surgery robots can help doctors reduce radiation and avoid carrying loads to perform surgery. In this paper, a robot system for liver interventional surgery is designed for the needs of embolization, a wireless control platform based on master-slave control mode is built, and the mapping relationship between the catheter guide wire and the robot motor drive is established. This paper proposes the ilter intelligent image recognition algorithm of YOLOv5-Kalman filter algorithm. It combines the motion state of the robot system, improves the recognition performance of the deep learning algorithm, ensures the real-time performance of the recognition algorithm and realizes the intraoperative image guide wire and Real-time tracking of catheter position. Experimental results show that the method has good accuracy and real-time performance, and can provide technical support for the safety warning, path planning, and autonomous navigation of surgical robots.