DG-TTA: Out-of-Domain Medical Image Segmentation Through Augmentation, Descriptor-Driven Domain Generalization, and Test-Time Adaptation
Sensors · 2025
DOI: 10.3390/s25175603 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, 2025. Verfasst von Christian Weihsbach und 3 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Medical Informatics, University of Lübeck, 23562 Lübeck, Germany. Hauptthemen laut MEDLINE: Image Processing, Computer-Assisted. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Bundesministerium für Forschung, Technologie und Raumfahrt, the German Federal Ministry of Education and Research (BMBF), the European Research Area Network on Cardiovascular Diseases (ERA-CVD).
Abstract (Originalsprache)
Applying pre-trained medical deep learning segmentation models to out-of-domain images often yields predictions of insufficient quality. In this study, we propose using a robust generalizing descriptor, along with augmentation, to enable domain-generalized pre-training and test-time adaptation, thereby achieving high-quality segmentation in unseen domains. In this study, five different publicly available datasets, including 3D CT and MRI images, are used to evaluate segmentation performance in out-of-domain scenarios. The settings include abdominal, spine, and cardiac imaging. Domain-generalized pre-training on source data is used to obtain the best initial performance in the target domain. We introduce a combination of the generalizing SSC descriptor and GIN intensity augmentation for optimal generalization. Segmentation results are subsequently optimized at test time, where we propose adapting the pre-trained models for every unseen scan using a consistency scheme with the augmentation–descriptor combination. The proposed generalized pre-training and subsequent test-time adaptation improve model performance significantly in CT to MRI cross-domain prediction for abdominal (+46.2 and +28.2 Dice), spine (+72.9), and cardiac (+14.2 and +55.7 Dice) scenarios (p < 0.001). Our method enables the optimal, independent use of source and target data, successfully bridging domain gaps with a compact and efficient methodology.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Institute of Medical Informatics, University of Lübeck, 23562 Lübeck, Germany.
- EchoScout GmbH, 23562 Lübeck, Germany.
- Dampsoft GmbH, 24351 Damp, Germany.
- Drägerwerk AG & Co. KGaA, 23558 Lübeck, Germany.
ORCID-Profile: Christian Weihsbach, Christian N Kruse, Alexander Bigalke, Mattias P Heinrich
Förderung
- Bundesministerium für Forschung, Technologie und Raumfahrt
- the German Federal Ministry of Education and Research (BMBF)
- the European Research Area Network on Cardiovascular Diseases (ERA-CVD)
Zitieren
APA
Weihsbach, C., Kruse, C. N., Bigalke, A., & Heinrich, M. P. (2025). DG-TTA: Out-of-Domain Medical Image Segmentation Through Augmentation, Descriptor-Driven Domain Generalization, and Test-Time Adaptation. Sensors. https://doi.org/10.3390/s25175603
Vancouver
Weihsbach C, Kruse CN, Bigalke A, Heinrich MP. DG-TTA: Out-of-Domain Medical Image Segmentation Through Augmentation, Descriptor-Driven Domain Generalization, and Test-Time Adaptation. Sensors. 2025. doi:10.3390/s25175603
BibTeX
@article{Weihsbach2025,
author = {Weihsbach, Christian and Kruse, Christian N and Bigalke, Alexander and Heinrich, Mattias P},
title = {DG-TTA: Out-of-Domain Medical Image Segmentation Through Augmentation, Descriptor-Driven Domain Generalization, and Test-Time Adaptation},
journal = {Sensors},
year = {2025},
doi = {10.3390/s25175603}
}
EUVIMED