Entropy Guided Unsupervised Domain Adaptation for Cross-Center Hip Cartilage Segmentation from MRI

Zeng, Guodong; Schmaranzer, Florian; Lerch, Till D.; Boschung, Adam; Zheng, Guoyan; Burger, Jürgen; Gerber, Kate; Tannast, Moritz; Siebenrock, K.; Kim, Young-Jo; Novais, Eduardo; Gerber, Nicolas (23 April 2021). Entropy Guided Unsupervised Domain Adaptation for Cross-Center Hip Cartilage Segmentation from MRI (Unpublished). In: Bern Data Science Day 2021. 23.04.2021. 10.5281/zenodo.4767469

[img]
Preview
Slideshow (Poster)
GD_Poster.pdf - Other
Available under License Creative Commons: Attribution (CC-BY).

Download (3MB) | Preview

Deep Convolutional Neural Networks have shown great success in various automatic medical image segmentation tasks, but testing on domain-shifted datasets (e.g. images obtained from different centers) can lead to severe performance losses. Our aim is to train a network which can realise cross-center hip MRI cartilage segmentation, without the need for additional time-consuming annotations on the target domain. In this abstract, we propose an entropy-guided unsupervised domain adaptation method and successfully demonstrate its application in the task of cross-center segmentation of the MRI hip cartilage, which is shown in below figure. Specifically, we first trained our model with supervised loss on the source domain, which enables low-entropy predictions on source-like images. Two discriminators were then used to minimize the gap between source and target domain with respect to the alignment of feature and entropy distribution. Compared to the results without adaptation with an average Dice of 46.46%, our method reports an average Dice value of 72.82%, which is quite close to the upper bound of 81.3% when the target annotations are directly included for training.

Item Type:

Conference or Workshop Item (Poster)

Division/Institute:

04 Faculty of Medicine > Faculty Institutions > sitem Center for Translational Medicine and Biomedical Entrepreneurship
04 Faculty of Medicine > Department of Orthopaedic, Plastic and Hand Surgery (DOPH) > Clinic of Orthopaedic Surgery
04 Faculty of Medicine > Department of Radiology, Neuroradiology and Nuclear Medicine (DRNN) > Institute of Diagnostic, Interventional and Paediatric Radiology
04 Faculty of Medicine > Department of Radiology, Neuroradiology and Nuclear Medicine (DRNN) > Institute of Diagnostic and Interventional Neuroradiology
08 Faculty of Science > School of Biomedical and Precision Engineering (SBPE)
08 Faculty of Science > School of Biomedical and Precision Engineering (SBPE) > Personalised Medicine

UniBE Contributor:

Zeng, Guodong, Schmaranzer, Florian, Lerch, Till, Boschung, Adam, Burger, Jürgen, Gerber, Kate, Siebenrock, Klaus-Arno, Gerber, Nicolas

Subjects:

600 Technology > 610 Medicine & health

Language:

English

Submitter:

Nicolas Gerber

Date Deposited:

06 Dec 2021 09:23

Last Modified:

24 Oct 2023 10:50

Publisher DOI:

10.5281/zenodo.4767469

Related URLs:

BORIS DOI:

10.48350/161451

URI:

https://boris.unibe.ch/id/eprint/161451

Actions (login required)

Edit item Edit item
Provide Feedback