Augmented Reality App on Tablet PC for Navigated Head and Neck Surgeries

Prof. Cattin, assistant-professor for Medical Image Analysis and Image-Guided Interventions at the University of Basel, Switzerland and his colleagues and clinical partners have recently published a paper about an Augmented Reality system for navigated head and neck surgery.

The research is motivated by the challenge of developing an „image-guided 3D navigation system that can be quickly brought in place, requires minimal training and is affordable.“ For this reason, the chosen hardware user interface is an off-the-shelf tablet PC serving as an Augmented Reality window that allows views through the tablet screen, through the patient’s skin onto patient’s anatomy.

Typically, tablet PCs show lower computational power than computers that are usually installed in surgical navigation systems. In addition to that, the tablet PC is equipped with only one camera that is used for inside-out tracking (camera pose is computed in real time) and hence correctly superimposing virtual objects on real objects. In contrast to this approach, state-of-the-art navigation systems using optical tracking locate marker trees attached to instruments and the patient usually with stereo camera setups. An extension of the tracking camera setup to multiple stereo views would solve the line of sight problem, when camera views are blocked for instance by instruments or surgical staff. This problem is even more present in the introduced single camera approach and is mentioned by Prof. Cattin in the discussion section of the paper.

These technical limitations of the chosen Augmented Reality enabling hardware require a novel tracking approach to map 3D markers attached to the anatomy of the patient and 2D projections of those markers in the camera images. This is a fundamental task to assess the 3D environment of the Augmented Reality scene and allow for accurate registration of real and virtual objects.

According to the authors, the reached mean error of 0.8mm between real and augmented object positions with a standard deviation of 1 mm is accurate enough to be “applicable for general guidance tasks during surgeries. Authors state that the proposed low-cost solution “could also find their applications in third-world countries or in remote areas”.

  • Adrian Schneider, Christian Baumberger, Mathias Griessen, Simon Pezold, Jörg Beinemann, Philipp Jürgens, Philippe C. Cattin: Landmark-Based Surgical Navigation. CLIP 2013: 57-64

Christoph Bichlmeier

Christoph Bichlmeier

Enthusiast of Augmented Reality for Medical Applications.

More Posts - Website

Follow Me:
FacebookLinkedIn