EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Simulation-to-real domain adaptation with teacher–student learning for endoscopic instrument segmentation

Manish Sahu, Anirban Mukhopadhyay, Stefan Zachow

International Journal of Computer Assisted Radiology and Surgery · 2021-05 · 16 (5) · 849-859

DOI: 10.1007/s11548-021-02383-4 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Computer Assisted Radiology and Surgery, online erschienen am 12.05.2021. Verfasst von Manish Sahu und 2 weiteren Autor:innen. Beteiligte Einrichtung: Zuse Institute Berlin (ZIB), Berlin, Germany. sahu@zib.de. Hauptthemen laut MEDLINE: Computer Simulation, Data Curation. Die Arbeit wurde bisher 30-mal zitiert (Crossref) und nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Bundesministerium für Bildung und Forschung.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

Segmentation of surgical instruments in endoscopic video streams is essential for automated surgical scene understanding and process modeling. However, relying on fully supervised deep learning for this task is challenging because manual annotation occupies valuable time of the clinical experts.

Methods

We introduce a teacher–student learning approach that learns jointly from annotated simulation data and unlabeled real data to tackle the challenges in simulation-to-real unsupervised domain adaptation for endoscopic image segmentation.

Results

Empirical results on three datasets highlight the effectiveness of the proposed framework over current approaches for the endoscopic instrument segmentation task. Additionally, we provide analysis of major factors affecting the performance on all datasets to highlight the strengths and failure modes of our approach.

Conclusions

We show that our proposed approach can successfully exploit the unlabeled real endoscopic video frames and improve generalization performance over pure simulation-based training and the previous state-of-the-art. This takes us one step closer to effective segmentation of surgical instrument in the annotation scarce setting.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Computer SimulationData CurationAlgorithmsArtifactsEndoscopyHumansImage Processing, Computer-AssistedLearningSoftwareStudentsVideo Recording

Stichwörter der Autor:innen

Consistency learningSelf-ensemblingSelf-supervisionSimulation-based learningSurgical instrument segmentationUnsupervised domain adaptation

Autor:innen und Institutionen

  • Zuse Institute Berlin (ZIB), Berlin, Germany. sahu@zib.de.
  • Department of Computer Science, TU Darmstadt, Darmstadt, Germany.
  • Zuse Institute Berlin (ZIB), Berlin, Germany.

Förderung

  • Bundesministerium für Bildung und Forschung

Zitieren

APA

Sahu, M., Mukhopadhyay, A., & Zachow, S. (2021). Simulation-to-real domain adaptation with teacher–student learning for endoscopic instrument segmentation. International Journal of Computer Assisted Radiology and Surgery, 16(5), 849-859. https://doi.org/10.1007/s11548-021-02383-4

Vancouver

Sahu M, Mukhopadhyay A, Zachow S. Simulation-to-real domain adaptation with teacher–student learning for endoscopic instrument segmentation. International Journal of Computer Assisted Radiology and Surgery. 2021;16(5):849-859. doi:10.1007/s11548-021-02383-4

BibTeX
@article{Sahu2021,
  author = {Sahu, Manish and Mukhopadhyay, Anirban and Zachow, Stefan},
  title = {Simulation-to-real domain adaptation with teacher–student learning for endoscopic instrument segmentation},
  journal = {International Journal of Computer Assisted Radiology and Surgery},
  year = {2021},
  volume = {16},
  number = {5},
  pages = {849--859},
  doi = {10.1007/s11548-021-02383-4}
}
Kostenlos starten →