Image-guided robotic bone fracture surgery is a medical procedure that automatically positions and controls a surgical robot for aligning fractured bones. In this robotic surgery, the key task is to find an accu-rate relationship between the robot and the bone, which is referred to as robot-patient registration. In the robot-patient registration, robot and bone structures often interfere mutually to hide important features which are used for a registration in X-ray images, thus leading to inaccurate registration. In order to solve this problem, an optimization method using two bidirectional X-ray images is pro-posed. Using alternative image information, undesired feature occlusion can be avoided. In addition, by the optimization based on the distance between features, registration error can be reduced. The experimental results show that the proposed robot-patient registration can improve the accuracy of the robot-patient registration.
Table Of Contents
Ⅰ. Introduction 1.1 Image-guided robotic surgery in bone fracture reduction 1 1.2 Conventional Robot-patient registration using 2D-3D registration 3 1.3 Problem 5 II. Method 2.1 Configuration of coordinate relationship between robot and patient 9 2.2 Obtaining the initial robot-patient registration 11 2.2.1 C-arm to robot registration using PnP algorithm 12 2.2.2 C-arm to bone registration using AR-based registration 14 2.3 Refining the initial robot-patient registration by optimization 17 2.3.1 Projection to get cost function 17 2.3.2 Optimization part 20 III. Results and Discussions IV. Conclusions
