Scalable Federated-Learning and Internet-of-Things enabled architecture for Chest Computer Tomography image classification. (September 2022)

• Record Type: Journal Article
• Authors: Dara, Suresh; Kanapala, Ambedkar; Babu, A. Ramesh; Dhamercherala, Swetha; Vidyarthi, Ankit; Agarwal, Ruchi
• Is Part Of: Computers & electrical engineering. Volume 102(2022)
• Publication Date: 2022-09
• Subjects: Federated Learning -- Big Data Analytics -- Internet of Medical Things (ioMT) -- COVID-19 -- Deep Learning -- Image classification; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2022.108266
• Languages: English
• ISSNs: 0045-7906

Abstract

Abstract: The recent proliferation of the Internet of Medical Things (IoMT), Federated Learning (FL), and Deep learning have opened new dimensions of research across the globe. This paper proposes the combined use of these paradigms to detect COVID-19 in Computer Tomography (CT) images. Initially, the framework collects the CT images at the various local hospital using IoMT and aggregated them in an Hadoop Distributed File system (HDFS) Spark big data framework for storage. Later, the proposed framework performs the model training in isolation with the trained parameters being sent to a centralized server for aggregation using federated Learning. The comprehensive experimentation is performed on three different COVID-19 databases to test the efficacy of the proposed work. The numerical investigation revealed that the proposed work outperforms existing techniques by a good margin. Also, the global server, when compared to the local server, achieves a 7.57% performance improvement in terms of accuracy and 3.33% in terms of Area Under Curve (AUC). Graphical abstract: Highlights: Designed the Federated Learning architecture to detect COVID-19 in CT images. Proposed a global ResNet-39 deep architecture for local servers. In experimentation the federated server outperforms ResNets at the node level. ResNet-39 witnessed an improvement of 7.57% in terms of Accuracy and 3.33% for AUC. The entire alliance has been deployed on Apache Spark.