Jie Zhou*, Yulong Shi*, Lin Qi, Xue Jiang, Shouliang Qi, Wei Qian. (*equal contribution)
Computer Methods and Programs in Biomedicine, 2025. IF: 4.8, JCR Q1
A3-DualUD proposes a source-free unsupervised domain adaptation method for cross-modality medical image segmentation. The approach combines anatomical anchor alignment with dual-path uncertainty denoising to address domain shift challenges without requiring access to source domain data during adaptation.
The framework employs two key components: (1) anatomical anchor alignment that leverages anatomical structure priors to guide domain alignment, and (2) dual-path uncertainty denoising that reduces noise in adaptation through complementary pathways. This combination enables effective source-free domain adaptation for medical image segmentation across different imaging modalities.
@article{zhou2025a3,
title={A3-DualUD: Source-free unsupervised domain adaptation via anatomical anchor alignment and dual-path uncertainty denoising for cross-modality medical image segmentation},
author={Zhou, Jie and Shi, Yulong and Qi, Lin and Jiang, Xue and Qi, Shouliang and Qian, Wei},
journal={Computer Methods and Programs in Biomedicine},
volume={271},
pages={109017},
year={2025},
publisher={Elsevier}
}