Niemeijer, Joshua und Ehrhardt, Jan und Kepp, Timo und Handels, Heinz und Schaefer, Jörg P. (2023) Overcoming the sensor delta for semantic segmentation in OCT images. In: Medical Imaging 2023: Computer-Aided Diagnosis. SPIE Medical Imaging, 2023-02-19 - 2023-02-24, San Diego. doi: 10.1117/12.2654187. ISBN 978-151066035-9. ISSN 1605-7422.
PDF
- Nur DLR-intern zugänglich
1MB |
Kurzfassung
The performance of a segmentation network optimized on data from a specific type of OCT sensor will decrease when applied to data from a different sensor. In this work, we deal with the research question of adapting models to data from an unlabeled new sensor with new properties in an unsupervised way. This challenge is known as unsupervised domain adaptation and can alleviate the need for costly manual annotation by radiologists. We show that one can strongly improve a model's result that was trained in a supervised way on the source OCT sensor domain on the target sensor domain. We do this by aligning the source and target domain distributions in the feature space through a semantic clustering method. Apart from the unsupervised domain adaptation we improved even the supervised training compared to the results in the RETOUCH challenge by employing a sophisticated training strategy. The RETOUCH challenge contains three different types of OCT scanners and provides annotations for the task of disease-related fluid classes.
elib-URL des Eintrags: | https://elib.dlr.de/198541/ | ||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Dokumentart: | Konferenzbeitrag (Vortrag) | ||||||||||||||||||||||||
Titel: | Overcoming the sensor delta for semantic segmentation in OCT images | ||||||||||||||||||||||||
Autoren: |
| ||||||||||||||||||||||||
Datum: | April 2023 | ||||||||||||||||||||||||
Erschienen in: | Medical Imaging 2023: Computer-Aided Diagnosis | ||||||||||||||||||||||||
Referierte Publikation: | Nein | ||||||||||||||||||||||||
Open Access: | Nein | ||||||||||||||||||||||||
Gold Open Access: | Nein | ||||||||||||||||||||||||
In SCOPUS: | Ja | ||||||||||||||||||||||||
In ISI Web of Science: | Nein | ||||||||||||||||||||||||
DOI: | 10.1117/12.2654187 | ||||||||||||||||||||||||
ISSN: | 1605-7422 | ||||||||||||||||||||||||
ISBN: | 978-151066035-9 | ||||||||||||||||||||||||
Status: | veröffentlicht | ||||||||||||||||||||||||
Stichwörter: | OCT, Segmentation, Unsupervised Learning, Domain Adaptation | ||||||||||||||||||||||||
Veranstaltungstitel: | SPIE Medical Imaging | ||||||||||||||||||||||||
Veranstaltungsort: | San Diego | ||||||||||||||||||||||||
Veranstaltungsart: | internationale Konferenz | ||||||||||||||||||||||||
Veranstaltungsbeginn: | 19 Februar 2023 | ||||||||||||||||||||||||
Veranstaltungsende: | 24 Februar 2023 | ||||||||||||||||||||||||
HGF - Forschungsbereich: | Luftfahrt, Raumfahrt und Verkehr | ||||||||||||||||||||||||
HGF - Programm: | Verkehr | ||||||||||||||||||||||||
HGF - Programmthema: | Straßenverkehr | ||||||||||||||||||||||||
DLR - Schwerpunkt: | Verkehr | ||||||||||||||||||||||||
DLR - Forschungsgebiet: | V ST Straßenverkehr | ||||||||||||||||||||||||
DLR - Teilgebiet (Projekt, Vorhaben): | V - KoKoVI - Koordinierter kooperativer Verkehr mit verteilter, lernender Intelligenz | ||||||||||||||||||||||||
Standort: | Braunschweig | ||||||||||||||||||||||||
Institute & Einrichtungen: | Institut für Verkehrssystemtechnik Institut für Verkehrssystemtechnik > Kooperative Systeme, BS | ||||||||||||||||||||||||
Hinterlegt von: | Niemeijer, Joshua | ||||||||||||||||||||||||
Hinterlegt am: | 05 Dez 2023 14:52 | ||||||||||||||||||||||||
Letzte Änderung: | 17 Okt 2024 08:18 |
Nur für Mitarbeiter des Archivs: Kontrollseite des Eintrags