Development of novel techniques for tracking surgical tools

点击翻译按钮获取中文摘要

This project aims to solve two issues we currently face in reliable localization and tracking of surgical tools, which is of paramount importance in orthopedic surgeries. To localize and track surgical tools, Think Surgical has developed a novel six-camera system to localize four markers (called fiducials) that are attached to each tool. Aim 1 focuses on reducing tracking error caused either by a transition of a fiducial marker out of the field of view of a tracking camera, or by an object (such as operator's hand) occluding one of the cameras within the field of view. We propose to fuse the physics of how the tools can move with sensory data of the six-camera system to address these issues. The physical prior information includes the fact that the velocity and acceleration of the tool cannot be infinity. Aim 2 focuses on detecting inaccurate localization of the surgical tools caused by an object partially covering a fiducial. Examples include a drop of blood or scrubs, which substantially increase the tracking error of the six-camera system. We propose two approaches to tackle this issue. First, we will build on the technique developed in Aim 1 to detect inaccurate sensory data. If all or few cameras localize a fiducial with a large error, the user will be alerted to inspect and clean that fiducial. Second, we will develop a machine learning technique to automatically classify each camera image of a fiducial as accurate or inaccurate. This research partnership can lead to development of novel technologies that enable more accurate and reliable localization of the surgical tools, potentially improving the outcome of orthopedic surgeries. It will also lead to training of two highly qualified personnel (HQP) working in close collaboration with a Canadian company.