Cross-Domain Distribution Alignment for Segmentation of Private Unannotated 3D Medical Images
ACM Transactions on Computing for Healthcare · 31.01.2027 · 8 (1) · 1-22
DOI: 10.1145/3822597 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Computing for Healthcare, online erschienen am 16.11.2026. Verfasst von Ruitong Sun und 1 weiteren Autor:innen. Die Arbeit nennt 72 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Manual annotation of 3D medical images for segmentation tasks is tedious and time-consuming. Moreover, data privacy limits the applicability of crowdsourcing to perform data annotation in medical domains. As a result, training deep neural networks for medical image segmentation can be challenging. We introduce a new source-free Unsupervised Domain Adaptation method to address this problem. Our idea is based on estimating the internally learned distribution of a relevant source domain by a base model and then sampling synthetic latent features that guide distribution alignment through a sliced Wasserstein loss. We demonstrate that our approach achieves competitive performance on real-world 3D medical datasets, including cross-modality cardiac segmentation (MMWHS) and cross-sequence brain tumor segmentation (BraTS), matching or approaching state-of-the-art results without requiring access to source data, which is a critical advantage for privacy-sensitive medical imaging applications.
Zitieren
APA
Sun, R., & Rostami, M. (2027). Cross-Domain Distribution Alignment for Segmentation of Private Unannotated 3D Medical Images. ACM Transactions on Computing for Healthcare, 8(1), 1-22. https://doi.org/10.1145/3822597
Vancouver
Sun R, Rostami M. Cross-Domain Distribution Alignment for Segmentation of Private Unannotated 3D Medical Images. ACM Transactions on Computing for Healthcare. 2027;8(1):1-22. doi:10.1145/3822597
BibTeX
@article{Sun2027,
author = {Sun, Ruitong and Rostami, Mohammad},
title = {Cross-Domain Distribution Alignment for Segmentation of Private Unannotated 3D Medical Images},
journal = {ACM Transactions on Computing for Healthcare},
year = {2027},
volume = {8},
number = {1},
pages = {1--22},
doi = {10.1145/3822597}
}
EUVIMED