TU Darmstadt / ULB / TUbiblio

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

Sahu, Manish ; Mukhopadhyay, Anirban ; Zachow, Stefan (2021)
Simulation-to-real domain adaptation with teacher–student learning for endoscopic instrument segmentation.
In: International Journal of Computer Assisted Radiology and Surgery, 16 (5)
doi: 10.1007/s11548-021-02383-4
Artikel, Bibliographie

Dies ist die neueste Version dieses Eintrags.

Kurzbeschreibung (Abstract)

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.

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.

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.

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.

Typ des Eintrags: Artikel
Erschienen: 2021
Autor(en): Sahu, Manish ; Mukhopadhyay, Anirban ; Zachow, Stefan
Art des Eintrags: Bibliographie
Titel: Simulation-to-real domain adaptation with teacher–student learning for endoscopic instrument segmentation
Sprache: Englisch
Publikationsjahr: 2021
Verlag: Springer
Titel der Zeitschrift, Zeitung oder Schriftenreihe: International Journal of Computer Assisted Radiology and Surgery
Jahrgang/Volume einer Zeitschrift: 16
(Heft-)Nummer: 5
DOI: 10.1007/s11548-021-02383-4
Zugehörige Links:
Kurzbeschreibung (Abstract):

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.

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.

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.

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.

Freie Schlagworte: Surgical instrument segmentation, Simulation-based learning, Self-supervision, Consistency learning, Self-ensembling, Unsupervised domain adaptation
Zusätzliche Informationen:

Erstveröffentlichung

Sachgruppe der Dewey Dezimalklassifikatin (DDC): 000 Allgemeines, Informatik, Informationswissenschaft > 004 Informatik
Fachbereich(e)/-gebiet(e): 20 Fachbereich Informatik
20 Fachbereich Informatik > Graphisch-Interaktive Systeme
Hinterlegungsdatum: 30 Sep 2024 11:24
Letzte Änderung: 30 Sep 2024 11:24
PPN:
Export:
Suche nach Titel in: TUfind oder in Google

Verfügbare Versionen dieses Eintrags

Frage zum Eintrag Frage zum Eintrag

Optionen (nur für Redakteure)
Redaktionelle Details anzeigen Redaktionelle Details anzeigen