Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation
In the wake of the Coronavirus disease (COVID-19) pandemic, chest computed tomography (CT) has become an invaluable component in the rapid and accurate detection of COVID-19. CT scans traditionally require manual inspections from medical professionals, which is expensive and tedious. With advancemen...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/169033 |
| _version_ | 1811683828124614656 |
|---|---|
| author | Feng, Yuanyi Luo, Yuemei Yang, Jianfei |
| author2 | School of Electrical and Electronic Engineering |
| author_facet | School of Electrical and Electronic Engineering Feng, Yuanyi Luo, Yuemei Yang, Jianfei |
| author_sort | Feng, Yuanyi |
| collection | NTU |
| description | In the wake of the Coronavirus disease (COVID-19) pandemic, chest computed tomography (CT) has become an invaluable component in the rapid and accurate detection of COVID-19. CT scans traditionally require manual inspections from medical professionals, which is expensive and tedious. With advancements in machine learning, deep neural networks have been applied to classify CT scans for efficient diagnosis. However, three challenges hinder this application of deep learning: (1) Domain shift across CT platforms and human subjects impedes the performance of neural networks in different hospitals. (2) Unsupervised Domain Adaptation (UDA), the traditional method to overcome domain shift, typically requires access to both source and target data. This is not realistic in COVID-19 diagnosis due to the sensitivity of medical data. The privacy of patients must be protected. (3) Data imbalance may exist between easy/hard samples and between data classes which can overwhelm the training of deep networks, causing degenerate models. To overcome these challenges, we propose a Cross-Platform Privacy-Preserving COVID-19 diagnosis network (CP 3 Net) that integrates domain adaptation, self-supervised learning, imbalanced label learning, and rotation classifier training into one synergistic framework. We also create a new CT benchmark by combining real-world datasets from multiple medical platforms to facilitate the cross-domain evaluation of our method. Through extensive experiments, we demonstrate that CP 3 Net outperforms many popular UDA methods and achieves state-of-the-art results in diagnosing COVID-19 using CT scans. |
| first_indexed | 2024-10-01T04:18:56Z |
| format | Journal Article |
| id | ntu-10356/169033 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T04:18:56Z |
| publishDate | 2023 |
| record_format | dspace |
| spelling | ntu-10356/1690332023-06-27T06:06:41Z Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation Feng, Yuanyi Luo, Yuemei Yang, Jianfei School of Electrical and Electronic Engineering Engineering::Electrical and electronic engineering Domain Adaptation Deep Learning In the wake of the Coronavirus disease (COVID-19) pandemic, chest computed tomography (CT) has become an invaluable component in the rapid and accurate detection of COVID-19. CT scans traditionally require manual inspections from medical professionals, which is expensive and tedious. With advancements in machine learning, deep neural networks have been applied to classify CT scans for efficient diagnosis. However, three challenges hinder this application of deep learning: (1) Domain shift across CT platforms and human subjects impedes the performance of neural networks in different hospitals. (2) Unsupervised Domain Adaptation (UDA), the traditional method to overcome domain shift, typically requires access to both source and target data. This is not realistic in COVID-19 diagnosis due to the sensitivity of medical data. The privacy of patients must be protected. (3) Data imbalance may exist between easy/hard samples and between data classes which can overwhelm the training of deep networks, causing degenerate models. To overcome these challenges, we propose a Cross-Platform Privacy-Preserving COVID-19 diagnosis network (CP 3 Net) that integrates domain adaptation, self-supervised learning, imbalanced label learning, and rotation classifier training into one synergistic framework. We also create a new CT benchmark by combining real-world datasets from multiple medical platforms to facilitate the cross-domain evaluation of our method. Through extensive experiments, we demonstrate that CP 3 Net outperforms many popular UDA methods and achieves state-of-the-art results in diagnosing COVID-19 using CT scans. Nanyang Technological University This work is supported by NTU Presidential Postdoctoral Fellowship, ‘‘Adaptive Multimodal Learning for Robust Sensing and Recognition in Smart Cities’’ project fund, at Nanyang Technological University, Singapore. 2023-06-27T06:06:40Z 2023-06-27T06:06:40Z 2023 Journal Article Feng, Y., Luo, Y. & Yang, J. (2023). Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation. Knowledge-Based Systems, 264, 110324-. https://dx.doi.org/10.1016/j.knosys.2023.110324 0950-7051 https://hdl.handle.net/10356/169033 10.1016/j.knosys.2023.110324 36713615 2-s2.0-85147848362 264 110324 en Knowledge-Based systems © 2023 Elsevier Ltd. All rights reserved. |
| spellingShingle | Engineering::Electrical and electronic engineering Domain Adaptation Deep Learning Feng, Yuanyi Luo, Yuemei Yang, Jianfei Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation |
| title | Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation |
| title_full | Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation |
| title_fullStr | Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation |
| title_full_unstemmed | Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation |
| title_short | Cross-platform privacy-preserving CT image COVID-19 diagnosis based on source-free domain adaptation |
| title_sort | cross platform privacy preserving ct image covid 19 diagnosis based on source free domain adaptation |
| topic | Engineering::Electrical and electronic engineering Domain Adaptation Deep Learning |
| url | https://hdl.handle.net/10356/169033 |
| work_keys_str_mv | AT fengyuanyi crossplatformprivacypreservingctimagecovid19diagnosisbasedonsourcefreedomainadaptation AT luoyuemei crossplatformprivacypreservingctimagecovid19diagnosisbasedonsourcefreedomainadaptation AT yangjianfei crossplatformprivacypreservingctimagecovid19diagnosisbasedonsourcefreedomainadaptation |