This is source code for the paper: "Federated Learning for COVID-19 Detection with Generative Adversarial Networks in Edge Cloud Computing", published at the IEEE Internet of Things Journal, Nov. 2021 (https://ieeexplore.ieee.org/abstract/document/9580478)
python >=3.5
tensorflow >= 2.6
pytorch >= 0.4
Install all required libraries and then is ready to run the code. It is recommended to run the standalone GAN code "COVID_GAN3.py" first. Then run the FL-GAN code "Server_COVID.py". Here, it is set for only 1 server and random 50 hospitals as hospital clients for training the COVID X-ray data. Then, use the Classifer Model for COVID detection. - "CNN_COVID_Classification.py" - "EFFICIENTNET_COVID_Classification.py" - "ENSEMBLE_B0B7_Classification.py"I used several open datasets.
- https://data.mendeley.com/datasets/rscbjbr9sj/2
- https://www.kaggle.com/datasets/pranavraikokte/covid19-image-dataset
- Etcetera.
The paper is available at https://arxiv.org/abs/2110.07136 and the authors using this code should cite as: @article{nguyen2021federatedcovid,
title={Federated learning for covid-19 detection with generative adversarial networks in edge cloud computing},
author={Nguyen, Dinh C and Ding, Ming and Pathirana, Pubudu N and Seneviratne, Aruna and Zomaya, Albert Y},
journal={IEEE Internet of Things Journal},
year={2021}, }