A YOLOv5 algorithm-based navigation method for minimally invasive pelvic acetabular surgery

Aiming at the difficult problem of minimally invasive pelvic acetabular surgical treatment which is difficult to place posterior column screws quickly and effectively, a minimally invasive pelvic acetabular surgical navigation system is designed. Based on the real intraoperative X-ray image data from the hospital's intraoperative C-arm ma-chine, the image dataset is constructed. The system is trained using the YOLOv5 net-work model and equipped with edge computing devices, which can process the X-ray images acquired by the intraoperative C-arm X-ray unit in real time, and with the adjustable template navigation, it can quickly generate effective surgical operation plans to assist the doctors to perform the clinical surgery. This design improves the success rate and accuracy of the operation, has low development cost, and can be used to carry out relevant operations even in ordinary county hospitals, providing a new processing method and idea for minimally invasive pelvic acetabular surgery.