Decentralized learning for medical image classification with prototypical contrastive network
Our proposed prototypical contrastive loss guides the local client's data distribution to align with the global distribution. Additionally, our method uses an approximate global prototype to address unbalanced dataset distribution across local clients by projecting all data onto a new balanced prototype space. Our model achieves state-of-the-art performance on the EyePACS, APTOS, IDRiD, and COVIDx datasets.
Authors
- Guanghui Yue (ORCID: https://orcid.org/0000-0001-9020-1735)
- Huazhen Lin
- Huanan Chen
- Weide Liu (ORCID: https://orcid.org/0000-0002-9855-4479)
- Yuanbing Shi
- Zhantao Cao
- Shuli Zhang (ORCID: https://orcid.org/0000-0001-8006-1134)
Institutions
- Agency for Science, Technology and Research (SG)
- Southwestern University of Finance and Economics (CN)
- Institute for Infocomm Research (SG)
- Science and Technology Department of Sichuan Province (CN)
- Shenzhen University Health Science Center (CN)
Publication Details
- Journal
- Medical Physics
- Published
- 2025-03-16
- DOI
- https://doi.org/10.1002/mp.17753
- Citations
- 1
- Primary Topic
- Retinal Imaging and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 1.36
Funders
- National Natural Science Foundation of China
- National Key Research and Development Program of China