Dynamic CNNs using uncertainty to overcome domain generalization for surgical instrument localization
Philipp M, Alperovich A, Gutt-Will M, Mathis A, Saur S, Raabe A, Mathis-Ullrich F (2022)
Publication Type: Conference contribution
Publication year: 2022
Publisher: Institute of Electrical and Electronics Engineers Inc.
Pages Range: 1727-1736
Conference Proceedings Title: Proceedings - 2022 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2022
Event location: Waikoloa, HI, USA
ISBN: 9781665409155
DOI: 10.1109/WACV51458.2022.00179
Abstract
Due to the limited amount of available annotated data in the medical field, domain generalization for applications in computer-assisted surgery is essential. Our work addresses this problem for the task of surgical instrument tip localization in neurosurgery, which is a classical step towards computer-assisted surgery. We propose an uncertainty-based CNN approach that dynamically selects the most relevant data source by incorporating its own uncertainty into the inference. In addition, the estimated uncertainty can visualize and easily explain the network's decision. Quantitative and qualitative evaluations show that our method outperforms state of the art approaches for large domain shifts and results are on-par for in-domain applications. Further increasing domain shifts by testing on different surgical disciplines, eye and laparoscopic surgeries, proves the generalization capabilities of the proposed method.
Authors with CRIS profile
Involved external institutions
Carl Zeiss Meditec AG
Germany (DE)
Carl Zeiss AG
Germany (DE)
Inselspital, Universitätsspital Bern
Switzerland (CH)
Karlsruhe Institute of Technology (KIT)
Germany (DE)
Germany (DE)
Carl Zeiss AG
Germany (DE)
Inselspital, Universitätsspital Bern
Switzerland (CH)
Karlsruhe Institute of Technology (KIT)
Germany (DE)
How to cite
APA:
Philipp, M., Alperovich, A., Gutt-Will, M., Mathis, A., Saur, S., Raabe, A., & Mathis-Ullrich, F. (2022). Dynamic CNNs using uncertainty to overcome domain generalization for surgical instrument localization. In Proceedings - 2022 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2022 (pp. 1727-1736). Waikoloa, HI, USA: Institute of Electrical and Electronics Engineers Inc..
MLA:
Philipp, Markus, et al. "Dynamic CNNs using uncertainty to overcome domain generalization for surgical instrument localization." Proceedings of the 22nd IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2022, Waikoloa, HI, USA Institute of Electrical and Electronics Engineers Inc., 2022. 1727-1736.
BibTeX: Download