• Medientyp: E-Artikel
  • Titel: Enhancing electromagnetic tracking accuracy in medical applications using pre-trained witness sensor distortion models
  • Beteiligte: Cavaliere, Marco; Cantillon-Murphy, Pádraig
  • Erschienen: Springer Science and Business Media LLC, 2023
  • Erschienen in: International Journal of Computer Assisted Radiology and Surgery
  • Sprache: Englisch
  • DOI: 10.1007/s11548-023-02994-z
  • ISSN: 1861-6429
  • Schlagwörter: Health Informatics ; Radiology, Nuclear Medicine and imaging ; General Medicine ; Surgery ; Computer Graphics and Computer-Aided Design ; Computer Science Applications ; Computer Vision and Pattern Recognition ; Biomedical Engineering
  • Entstehung:
  • Anmerkungen:
  • Beschreibung: <jats:title>Abstract</jats:title><jats:sec> <jats:title>Purpose</jats:title> <jats:p>Electromagnetic tracking (EMT) accuracy is affected by the presence of surrounding metallic materials. In this work, we propose measuring the magnetic field's variation due to distortion at a witness position to localise the instrument causing distortion based on a pre-trained model and without additional sensors attached to it. </jats:p> </jats:sec><jats:sec> <jats:title>Methods</jats:title> <jats:p>Two experiments were performed to demonstrate possible applications of the technique proposed. In the first case, the distortion introduced by an ultrasound (US) probe was characterised and subsequently used to track the probe position on a line. In the second application, the measurement was used to estimate the distance of an interventional fluoroscopy C-arm machine and apply the correct compensation model. </jats:p> </jats:sec><jats:sec> <jats:title>Results</jats:title> <jats:p>Tracking of the US probe using the proposed method was demonstrated with millimetric accuracy. The distortion created by the C-arm caused errors in the order of centimetres, which were reduced to 1.52 mm RMS after compensation. </jats:p> </jats:sec><jats:sec> <jats:title>Conclusions</jats:title> <jats:p>The distortion profile associated with medical equipment was pre-characterised and used in applications such as object tracking and error compensation map selection. In the current study, the movement was limited to one degree of freedom (1 DOF) and simple analytical functions were used to model the magnetic distortion. Future work will explore advanced AI models to extend the method to 6 DOF tracking using multiple witness sensors.</jats:p> </jats:sec>