Multi-target Attachment for Surgical Instrument Tracking

Abstract

The pose estimation of a surgical instrument is a common problem in the new needs of medical science. Many instrument tracking methods use markers with a known geometry that allows for solving the instrument pose as detected by a camera. However, marker occlusion happens, and it hinders correct pose estimation. In this work, we propose an adaptable multi-target attachment with ArUco markers to solve occlusion problems on tracking a medical instrument like an ultrasound probe or a scalpel. Our multi-target system allows for precise and redundant real-time pose estimation implemented in OpenCV. Encouraging results show that the multi-target device may prove useful in the clinical setting. © 2021, Springer Nature Switzerland AG.