AUTHOR=Liu Xiaofeng , Yoo Chaehwa , Xing Fangxu , Kuo C.-C. Jay , El Fakhri Georges , Kang Je-Won , Woo Jonghye 

TITLE=Unsupervised Black-Box Model Domain Adaptation for Brain Tumor Segmentation

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 16 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2022.837646

DOI=10.3389/fnins.2022.837646

ISSN=1662-453X

ABSTRACT=Unsupervised domain adaptation (UDA) is an emerging technique that enables to transfer domain knowledge learned from a labeled source domain to unlabeled target domains to cope with the difficulty of labeling in new domains. The majority of prior work has relied on both source and target domain data for adaptation. However, because of the privacy concerns over leakage of sensitive information contained in patient data, it is often challenging to share the data and their labels in the source domain and trained model parameters in cross-center collaborations. To address this issue, we propose a practical framework for UDA with a black-box segmentation model trained in the source domain only, without relying on original source data or a white-box source model in which the network parameters are accessible. In particular, we propose a knowledge distillation scheme with exponential mixup decay to gradually learn target-specific representations. Additionally, we regularize the confidence of the labels in the target domain via unsupervised entropy minimization, leading to performance gain over UDA without entropy minimization. We extensively validated our framework on several datasets and deep learning backbones, demonstrating the potential of our framework to be applied in challenging yet realistic clinical settings.