Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Jul 24;20(7):e0326579.
doi: 10.1371/journal.pone.0326579. eCollection 2025.

FedGAN: Federated diabetic retinopathy image generation

Hassan Kamran  1 , Syed Jawad Hussain  1 , Sohaib Latif  2 , Imtiaz Ali Soomro  1 , Mrim M Alnfiai  3 , Nouf Nawar Alotaibi  4
Affiliations

FedGAN: Federated diabetic retinopathy image generation

Hassan Kamran et al. PLoS One. .

Abstract

Deep learning models for diagnostic applications require large amounts of sensitive patient data, raising privacy concerns under centralized training paradigms. We propose FedGAN, a federated learning framework for synthetic medical image generation that combines Generative Adversarial Networks (GANs) with cross-silo federated learning. Our approach pretrains a DCGAN on abdominal CT scans and fine-tunes it collaboratively across clinical silos using diabetic retinopathy datasets. By federating the GAN's discriminator and generator via the Federated Averaging (FedAvg) algorithm, FedGAN generates high-quality synthetic retinal images while complying with HIPAA and GDPR. Experiments demonstrate that FedGAN achieves a realism score of 0.43 (measured by a centralized discriminator). This work bridges data scarcity and privacy challenges in medical AI, enabling secure collaboration across institutions.

PubMed Disclaimer

Conflict of interest statement

There is no competing interest.

Figures

Fig 1
Fig 1. Sample RSNA dataset.
Fig 2
Fig 2. DCGAN generator.
Fig 3
Fig 3. DCGAN discriminator.
Fig 4
Fig 4. Unfederated loss.
Fig 5
Fig 5. Realism score comparison.
Fig 6
Fig 6. Generated samples from experiment with 3 clients.
Fig 7
Fig 7. Generated samples from experiment with 5 clients.
Fig 8
Fig 8. Generated Samples from experiment with 7 clients.
Fig 9
Fig 9. Generated samples from experiment with 10 clients.
Fig 10
Fig 10. Sample diabetes retinopathy dataset.
Fig 11
Fig 11. Preprocessed images.
Fig 12
Fig 12. Output of epoch from pre-training step.
Fig 13
Fig 13. Unfederated learning training samples.
Fig 14
Fig 14. Federated learning results.
Fig 15
Fig 15. Federated learning discriminator loss.
Fig 16
Fig 16. Federated learning discriminator loss variance.
Fig 17
Fig 17. Federated learning generator loss.
Fig 18
Fig 18. Federated learning generator loss variance.
See this image and copyright information in PMC

References

    1. Ho-Phuoc T. CIFAR10 to compare visual recognition performance between deep neural networks and humans. 2019. doi: 10.48550/arXiv.1811.07270 - DOI
    1. Hunter P. The advent of AI and deep learning in diagnostics and imaging: Machine learning systems have potential to improve diagnostics in healthcare and imaging systems in research. EMBO Rep. 2019;20(7):e48559. doi: 10.15252/embr.201948559 - DOI - PMC - PubMed
    1. Economics of artificial intelligence in healthcare: diagnosis vs. treatment. Healthcare. https://www.mdpi.com/2227-9032/10/12/2493. Accessed 2024 August 25. - PMC - PubMed
    1. Roppelt JS, Kanbach DK, Kraus S. Artificial intelligence in healthcare institutions: A systematic literature review on influencing factors. Technol Soc. 2024;76:102443. doi: 10.1016/j.techsoc.2023.102443 - DOI
    1. Rights O. Summary of the HIPAA privacy rule. https://www.hhs.gov/hipaa/for-professionals/privacy/laws-regulations/ind.... Accessed 2024 August 18.