Unsupervised domain adaptation (UDA) has been widely used to transfer knowledge from a labeled source domain to an unlabeled target domain to counter the difficulty of labeling in a new domain. The training of conventional solutions usually relies on the existence of both source and target domain data. However, privacy of the large-scale and well-labeled data in the source domain and trained model parameters can become the major concern of cross center/domain collaborations. In this work, to address this, we propose a practical solution to UDA for segmentation with a black-box segmentation model trained in the source domain only, rather than original source data or a white-box source model. Specifically, we resort to a knowledge distillation scheme with exponential mixup decay (EMD) to gradually learn target-specific representations. In addition, unsupervised entropy minimization is further applied to regularization of the target domain confidence. We evaluated our framework on the BraTS 2018 database, achieving performance on par with white-box source model adaptation approaches.
|Title of host publication||Medical Imaging 2022|
|Subtitle of host publication||Image Processing|
|Editors||Olivier Colliot, Ivana Isgum, Bennett A. Landman, Murray H. Loew|
|State||Published - 2022|
|Event||Medical Imaging 2022: Image Processing - Virtual, Online|
Duration: 21 Mar 2021 → 27 Mar 2021
|Name||Progress in Biomedical Optics and Imaging - Proceedings of SPIE|
|Conference||Medical Imaging 2022: Image Processing|
|Period||21/03/21 → 27/03/21|
Bibliographical noteFunding Information:
This work is partially supported by NIH R01DC018511 and P41EB022544.
© 2022 SPIE.
- Black-box source model
- Brain MR image segmentation
- Unsupervised domain adaptation