Volltext-Downloads (blau) und Frontdoor-Views (grau)

Binary endoscope localization from laparoscopic video as virtual sensor for automatic workflow detection

  • This paper contributes to the automatic detection of perioperative workflow by developing a binary endoscope localization. Automated situation recognition in the context of an intelligent operating room requires the automatic conversion of low level cues into more abstract high level information. Imagery from a laparoscope delivers rich content that is easy to obtain but hard to process. We introduce a system which detects if the endoscope's distal tip is inside or outsiede the patient based on the endoscope video. This information can be used as one parameter in a situation recognition pipeline. Our localization performs in real-time at a video resolution of 1280x720 and 5-fold cross validation yields mean F1-scores of up to 0,94 on videos of 7 laparoscopies.

Export metadata

Additional Services

Share in Twitter Search Google Scholar


Author of HS ReutlingenScheytt, Josia; Wiemuth, Markus; Burgert, Oliver
Erschienen in:CURAC 2017 - Tagungsband : 16. Jahrestagung der Deutschen Gesellschaft für Computer- und Roboterassistierte Chirurgie (CURAC) : 5.-7. Oktober 2017, Hannover
Publisher:Deutsche Gesellschaft für Computer- und Roboterassistierte Chirurgie e. V.
Place of publication:Stuttgart
Editor:Jessica Burgner-Kahrs
Document Type:Conference proceeding
Publication year:2017
Tag:endoscope localization; laparoscopic surgery; situation recognition
Page Number:6
First Page:191
Last Page:196
PPN:Im Katalog der Hochschule Reutlingen ansehen
DDC classes:610 Medizin, Gesundheit
Open access?:Ja
Licence (German):License Logo  Open Access