Federated Semi-Supervised Multi-Task Learning to Detect COVID-19 and Lungs Segmentation Marking Using Chest Radiography Images and Raspberry Pi Devices: An Internet of Medical Things Application

Abstract

Internet of Medical Things (IoMT) provides an excellent opportunity to investigate better automatic medical decision support tools with the effective integration of various medical equipment and associated data. This study explores two such medical decision-making tasks, namely COVID-19 detection and lung area segmentation detection, using chest radiography images. We also explore different cutting-edge machine learning techniques, such as federated learning, semi-supervised learning, transfer learning, and multi-task learning to explore the issue. To analyze the applicability of computationally less capable edge devices in the IoMT system, we report the results using Raspberry Pi devices as accuracy, precision, recall, Fscore for COVID-19 detection, and average dice score for lung segmentation detection tasks. We also publish the results obtained through server-centric simulation for comparison. The results show that Raspberry Pi-centric devices provide better performance in lung segmentation detection, and server-centric experiments provide better results in COVID-19 detection. We also discuss the IoMT application-centric settings, utilizing medical data and decision support systems, and posit that such a system could benefit all the stakeholders in the IoMT domain.

Telemedicine and virtual respiratory care in the era of COVID-19

“…[164] AI techniques have been applied to COVID-related epidemiology, therapeutics (drug discovery), clinical research (predictive modelling), social and behavioural studies. [165] Specific examples include development of automated tools for interpreting CT scans, chest X-rays and lung ultrasound diagnostics, [166][167][168][169][170] though clinical trials to establish the benefits and safety of image analysis in clinical practice are awaited. [171]…”

Federated Learning for Medical Applications: A Taxonomy, Current Trends, Challenges, and Future Research Directions

“…The authors in [160] introduced a collaborative FL framework that allows multiple medical institutions to use DL to screen COVID-19 from chest X-ray images without sharing/exchanging the patient data. Similarly, there have been several studies that propose the use of FL for COVID-19 X-ray data training and deploy experiments to verify the FL effectiveness [161,162,163,164,165,166,167,168,169,58,170,171,172,173].…”

A survey of COVID-19 detection and prediction approaches using mobile devices, AI, and telemedicine

“…Third, patients' data privacy is imperative to be secured. It will need engineering effort to integrate technology such as federated learning, and blockchain (a back-linked database with cryptographic protocols) to achieve the goal (Sufian et al, 2020;Alam and Rahmani, 2021;Hassan et al, 2021;Tahiliani et al, 2021). With telemedicine and related IoT devices fully equipped, health systems can be more prepared for COVID-19 and other future pandemics (Gupta et al, 2021).…”

“…This architecture of learning has been a focus for COVID-19 infection prediction due to the fact that many medical devices such as wrist bands are edge devices which can transmit data to smart phone and process the information (Hassan et al, 2021). Federated learning has been utilized for both traditional learning as well as transfer learning (Sufian et al, 2020;Alam and Rahmani, 2021). To provide further security protection and scalability, research has been conducted in integrating federated learning systems into the cloud.…”