Work done at Department of Computer Science, Technischer Universität Darmstadt, Darmstadt, Germany.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Accurate estimation of the position and orientation (pose) of surgical instruments is crucial for delicate minimally invasive temporal bone surgery. Current techniques lack in accuracy and/or line-of-sight constraints (conventional tracking systems) or expose the patient to prohibitive ionizing radiation (intra-operative CT). A possible solution is to capture the instrument with a c-arm at irregular intervals and recover the pose from the image.
i3PosNet infers the position and orientation of instruments from images using a pose estimation network. Said framework considers localized patches and outputs pseudo-landmarks. The pose is reconstructed from pseudo-landmarks by geometric considerations.
We show i3PosNet reaches errors \(<\,0.05\) mm. It outperforms conventional image registration-based approaches reducing average and maximum errors by at least two thirds. i3PosNet trained on synthetic images generalizes to real X-rays without any further adaptation.
The translation of deep learning-based methods to surgical applications is difficult, because large representative datasets for training and testing are not available. This work empirically shows sub-millimeter pose estimation trained solely based on synthetic training data.