Preliminary Design and Evaluation of an Interfacing Mechanism for Maneuvering Virtual Minimally Invasive Surgical Instruments

Abstract

Augmenting the motion of virtual surgical instruments onto a minimally invasive surgical field acts as a visual cue for the operating surgeon. In this work we propose an interfacing mechanism to provide input for maneuvering such virtual surgical instruments. Specifically, an interface in the form of a 3D-printed dodecahedron pen with attached binary squared planar markers is employed. The proposed tracking mechanism computes the pose of the interface from a realtime video feed acquired from a camera. The system provides accurate pose estimation with mean errors of $0.27 \pm 0.06$ mm in translation and $0.37 \pm 0.04$ degrees in rotation. The object pose estimation takes ~6 ms. Utilized Azure Kinect camera with frame rate of 30 FPS and 1280 x 720 image resolution video, the tracking speed of the proposed system is ~25 FPS. The easy to integrate, cost effective setup makes the interfacing mechanism particularly suitable for remote surgical tele-mentoring applications.