Computer Vision News - March 2023

31 Ganymed Robotics The team has used different methods to verify and validate the efficacy of their algorithm across a range of data sources, including simulated and phantom data using synthetic plastic bones and rigorous testing on cadavers under laboratory conditions . The cadaver testing used trackers or landmarks to obtain the ground truth as a benchmark for evaluating the algorithm’s performance. Working with hard structures like bones Preoperative data comes from a 3D model of the bone extracted from a CT scan. “ We begin by detecting the bone in the images and merge the preoperative and intraoperative data to locate it, ” Marion explains. “ Once we know the orientation and position of the bone, we know it won’t move with respect to the robot. Then the robotic part of the system leads on tracking and guides the surgeon with precision and accuracy. ”