gms | German Medical Science

72. Jahrestagung der Deutschen Gesellschaft für Neurochirurgie (DGNC)
Joint Meeting mit der Polnischen Gesellschaft für Neurochirurgie

Deutsche Gesellschaft für Neurochirurgie (DGNC) e. V.

06.06. - 09.06.2021

Reconstruction of real 3D models from endoscopic images using photogrammetry

Rekonstruktion von 3D Models aus endoskopischen Bildern mit Fotogrammmetrie

Meeting Abstract

  • presenting/speaker László Bárány - Semmelweis-Universität, Anatomisches Institut, Budapest, Ungarn
  • Oliver Ganslandt - Klinikum Stuttgart, Katharinenhospital, Neurochirurgische Klinik, Stuttgart, Deutschland
  • Michael Buchfelder - Friedrich-Alexander University (FAU) Erlangen-Nürnberg, Neurochirurgische Klinik, Erlangen, Deutschland
  • Péter Kurucz - Klinikum Stuttgart, Katharinenhospital, Neurochirurgische Klinik, Stuttgart, Deutschland; Friedrich-Alexander University (FAU) Erlangen-Nürnberg, Neurochirurgische Klinik, Erlangen, Deutschland

Deutsche Gesellschaft für Neurochirurgie. 72. Jahrestagung der Deutschen Gesellschaft für Neurochirurgie (DGNC), Joint Meeting mit der Polnischen Gesellschaft für Neurochirurgie. sine loco [digital], 06.-09.06.2021. Düsseldorf: German Medical Science GMS Publishing House; 2021. DocP085

doi: 10.3205/21dgnc373, urn:nbn:de:0183-21dgnc3732

Published: June 4, 2021

© 2021 Bárány et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. See license information at http://creativecommons.org/licenses/by/4.0/.


Outline

Text

Objective: The comprehensive visuospatial understanding of neuroanatomy is crucial for neurosurgeons. Cadavers represent the gold standard of learning; however, they are very expensive, and their accessibility is also limited. To overcome the limitations associated with cadavers, different computer-assisted methods were developed in the last years to facilitate neurosurgical education. One of them is based on photogrammetry, which means the reconstruction of three-dimensional volumetric models from two-dimensional images. The aim of our study was to evaluate the feasibility of creating freely explorable 3D models from endoscopic images and short videos using photogrammetric approach.

Methods: 10 in formalin-fixed human cadaveric heads were involved in this study. The cadavers were donated with educational and research purposes. The opening of the skull and the dura mater was performed according to the classical supraorbital and retrosigmoid approaches. The anatomical structures of the suprasellar region as well as of the cerebellopontine angle were exposed and recorded (in 8 cases) or photographed (in 5 cases) from different angles and distances with a 0-degree rigid endoscope (Aesculap GmbH, Tuttlingen, Germany) mounted on a Canon EOS 5D Mark II body (Canon Inc, Tokio, Japan). To eliminate the optical distortion caused by the endoscope lens, a chessboard pattern was photographed with the same setup. The distortion parameters were estimated using the OpenCV open source software. The images were extracted from the videos before the photogrammetric processing. The three-dimensional volumetric models were reconstructed using the Agisoft Metashape software (Standard Edition, v. 1.6.1)

Results: We have successfully created freely explorable three dimensional models from the endoscopic images and videos in all cases. In 2 cases we were able to reconstruct multiplanar models which allowed the exploration of the deeper structures between the superficial ones. Comparing the two different inputs, the extracted images were also suitable for photogrammetric processing despite their significantly lower resolution.

Conclusion: Photogrammetry is a suitable method to creating volumetric three-dimensional models from endoscopic images for educational purposes. The snapshots extracted from videos are not inferior compared to high quality photographs. The intraoperative application of this technique should be also considered.