Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh Scans
Semantic segmentation over three-dimensional (3D) intra-oral mesh scans (IOS) is an essential step in modern digital dentistry. Many existing methods usually rely on a limited number of labeled samples as annotating IOS scans is time consuming, while a large-scale dataset of IOS is not yet publicly...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-06-01
|
| Series: | Frontiers in Communications and Networks |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/frcmn.2022.907388/full |
| _version_ | 1811333943552638976 |
|---|---|
| author | Songshang Liu Howard H. Yang Yiqi Tao Yang Feng Yang Feng Jin Hao Zuozhu Liu |
| author_facet | Songshang Liu Howard H. Yang Yiqi Tao Yang Feng Yang Feng Jin Hao Zuozhu Liu |
| author_sort | Songshang Liu |
| collection | DOAJ |
| description | Semantic segmentation over three-dimensional (3D) intra-oral mesh scans (IOS) is an essential step in modern digital dentistry. Many existing methods usually rely on a limited number of labeled samples as annotating IOS scans is time consuming, while a large-scale dataset of IOS is not yet publicly available due to privacy and regulatory concerns. Moreover, the local data heterogeneity would cause serious performance degradation if we follow the conventional learning paradigms to train local models in individual institutions. In this study, we propose the FedTSeg framework, a federated 3D tooth segmentation framework with a deep graph convolutional neural network, to resolve the 3D tooth segmentation task while alleviating data privacy issues. Moreover, we adopt a general privacy-preserving mechanism with homomorphic encryption to prevent information leakage during parameter exchange between the central server and local clients. Extensive experiments demonstrate that both the local and global models trained with the FedTSeg framework can significantly outperform models trained with the conventional paradigm in terms of the mean intersection over union, dice coefficient, and accuracy metrics. The FedTSeg framework can achieve better performance under imbalanced data distributions with different numbers of clients, and its overall performance is on par with the central model trained with the full dataset aggregated from all distributed clients. The data privacy during parameter exchange of FedTSeg is further enhanced with a homomorphic encryption process. Our work presents the first attempts of federated learning for 3D tooth segmentation, demonstrating its strong potential in challenging federated 3D medical image analysis in multi-centric settings. |
| first_indexed | 2024-04-13T17:00:21Z |
| format | Article |
| id | doaj.art-4bcf9fc7e1854462a6a9bbff38725039 |
| institution | Directory Open Access Journal |
| issn | 2673-530X |
| language | English |
| last_indexed | 2024-04-13T17:00:21Z |
| publishDate | 2022-06-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Communications and Networks |
| spelling | doaj.art-4bcf9fc7e1854462a6a9bbff387250392022-12-22T02:38:41ZengFrontiers Media S.A.Frontiers in Communications and Networks2673-530X2022-06-01310.3389/frcmn.2022.907388907388Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh ScansSongshang Liu0Howard H. Yang1Yiqi Tao2Yang Feng3Yang Feng4Jin Hao5Zuozhu Liu6ZJU-UIUC Institute, ZJU-Angelalign R&D Center for Intelligent Healthcare, Zhejiang University, Hangzhou, ChinaZJU-UIUC Institute, ZJU-Angelalign R&D Center for Intelligent Healthcare, Zhejiang University, Hangzhou, ChinaZJU-UIUC Institute, ZJU-Angelalign R&D Center for Intelligent Healthcare, Zhejiang University, Hangzhou, ChinaZJU-UIUC Institute, ZJU-Angelalign R&D Center for Intelligent Healthcare, Zhejiang University, Hangzhou, ChinaAngelalign Inc., Shanghai, ChinaHarvard School of Dental Medicine, Harvard University, Boston, MA, United StatesZJU-UIUC Institute, ZJU-Angelalign R&D Center for Intelligent Healthcare, Zhejiang University, Hangzhou, ChinaSemantic segmentation over three-dimensional (3D) intra-oral mesh scans (IOS) is an essential step in modern digital dentistry. Many existing methods usually rely on a limited number of labeled samples as annotating IOS scans is time consuming, while a large-scale dataset of IOS is not yet publicly available due to privacy and regulatory concerns. Moreover, the local data heterogeneity would cause serious performance degradation if we follow the conventional learning paradigms to train local models in individual institutions. In this study, we propose the FedTSeg framework, a federated 3D tooth segmentation framework with a deep graph convolutional neural network, to resolve the 3D tooth segmentation task while alleviating data privacy issues. Moreover, we adopt a general privacy-preserving mechanism with homomorphic encryption to prevent information leakage during parameter exchange between the central server and local clients. Extensive experiments demonstrate that both the local and global models trained with the FedTSeg framework can significantly outperform models trained with the conventional paradigm in terms of the mean intersection over union, dice coefficient, and accuracy metrics. The FedTSeg framework can achieve better performance under imbalanced data distributions with different numbers of clients, and its overall performance is on par with the central model trained with the full dataset aggregated from all distributed clients. The data privacy during parameter exchange of FedTSeg is further enhanced with a homomorphic encryption process. Our work presents the first attempts of federated learning for 3D tooth segmentation, demonstrating its strong potential in challenging federated 3D medical image analysis in multi-centric settings.https://www.frontiersin.org/articles/10.3389/frcmn.2022.907388/fullfederated learningmedical image analysissemantic segmentationhomomorphic encryptiontooth segmentation |
| spellingShingle | Songshang Liu Howard H. Yang Yiqi Tao Yang Feng Yang Feng Jin Hao Zuozhu Liu Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh Scans Frontiers in Communications and Networks federated learning medical image analysis semantic segmentation homomorphic encryption tooth segmentation |
| title | Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh Scans |
| title_full | Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh Scans |
| title_fullStr | Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh Scans |
| title_full_unstemmed | Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh Scans |
| title_short | Privacy-Preserved Federated Learning for 3D Tooth Segmentation in Intra-Oral Mesh Scans |
| title_sort | privacy preserved federated learning for 3d tooth segmentation in intra oral mesh scans |
| topic | federated learning medical image analysis semantic segmentation homomorphic encryption tooth segmentation |
| url | https://www.frontiersin.org/articles/10.3389/frcmn.2022.907388/full |
| work_keys_str_mv | AT songshangliu privacypreservedfederatedlearningfor3dtoothsegmentationinintraoralmeshscans AT howardhyang privacypreservedfederatedlearningfor3dtoothsegmentationinintraoralmeshscans AT yiqitao privacypreservedfederatedlearningfor3dtoothsegmentationinintraoralmeshscans AT yangfeng privacypreservedfederatedlearningfor3dtoothsegmentationinintraoralmeshscans AT yangfeng privacypreservedfederatedlearningfor3dtoothsegmentationinintraoralmeshscans AT jinhao privacypreservedfederatedlearningfor3dtoothsegmentationinintraoralmeshscans AT zuozhuliu privacypreservedfederatedlearningfor3dtoothsegmentationinintraoralmeshscans |