Home / Current Issue / Paper 1715262
Federated Learning Framework for Integrating Multi-Device IoT Data in Healthcare Applications
Subject area: Science,Engineering and Technology · Area of research: AI And ML
Abstract
There has been an increase in the volume of distributed, heterogeneous healthcare data with the advent of Internet of Things (IoT) technology-based medical devices. However, aggregating large volumes of sensitive patient information at a central server poses several privacy, security, and regulatory hurdles. This paper proposes a decentralized Federated Learning (FL) architecture that can accommodate multi-device IoT-based medical data integration without sharing sensitive patient information between healthcare institutions. Each institution maintains their unique model, which can be used for training on their premises, and then the models are integrated into a single federated model for collective use. A novel framework is proposed, which can accommodate collaborative training between several nodes of healthcare institutions, along with local processing of multimodal data obtained from IoT medical devices. The multimodal deep neural network architecture has been used to combine different types of input, including medical images and structured demographic or clinical attributes, through parallel feature extraction branches followed by feature-level fusion. Each healthcare institution has a unique model, which is trained on their respective data, and then the model parameters are sent to a central server for aggregation through the Federated Averaging (FedAvg) algorithm. Cross-validation of the models has also been done to check the robustness of the models, which involves training on several nodes and then validation on an external dataset. The results obtained through the federated multimodal framework indicate high classification accuracy, thus ensuring the viability of a secure, scalable, and privacy-preserving collaborative intelligence system for IoT-based healthcare environments.
References
[1] H. Lee, Y. J. Chai, H. Joo, et al., “Federated learning for thyroid ultrasound image analysis to protect personal information: Validation study in a real health care environment,” JMIR Medical Informatics, vol. 9, no. 5, e25869, 2021.
[2] Z. Xiang, et al., “Federated learning via multi-attention guided UNet for thyroid nodule segmentation,” Neurocomputing, 2025.
[3] M. Savelonas, et al., “An overview of AI-guided thyroid ultrasound image analysis: Three-dimensional imaging, Doppler, federated learning, and future directions,” Knowledge, vol. 9, no. 10, 2025.
[4] M. Moshawrab, et al., “Reviewing federated machine learning and its use in disease prediction,” Sensors, vol. 23, no. 4, p. 2112, 2023.
[5] S. Nazir, et al., “Federated learning for medical image analysis with deep neural networks: A survey,” Diagnostics, vol. 13, no. 9, p. 1532, 2023.
[6] “Privacy-preserving federated learning for thyroid disease detection: A comparative study on feature selection and class balancing,” ResearchGate Preprint, 2025.
[7] M. Hu, et al., “Development and preliminary validation of a machine learning approach to diagnose thyroid dysfunction from routine laboratory tests,” 2022.
[8] J. Y. Noh, et al., “An evaluation of the efficacy of machine learning in predicting thyrotoxicosis and hypothyroidism,” Internal Medicine, 2024.
[9] K. Shin, et al., “A machine learning-assisted system to predict the occurrence of thyrotoxicosis using heart-rate data from wearable devices,” 2023.
[10] R. Chaganti, et al., “Thyroid disease prediction using selective features and machine learning,” 2022.
[11] S. Ji, et al., “SSC: A self-stack ensemble model for thyroid disease prediction,” PLOS ONE, 2024.
[12] C. Chantasartrassamee, et al., “AI-enhanced infrared thermography for thyroid nodules,” 2024.
[13] N. Zhang, et al., “Multi-modal thyroid nodule classification using ultrasound and infrared imaging,” BMC Bioinformatics, 2023.
How to cite this paper
@article{1715262,
author = {Ragul K B, Kumaran S, Arshad Ahmed K, Rohini C},
title = {Federated Learning Framework for Integrating Multi-Device IoT Data in Healthcare Applications},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {9},
pages = {1471-1479},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1715262.pdf},
abstract = {There has been an increase in the volume of distributed, heterogeneous healthcare data with the advent of Internet of Things (IoT) technology-based medical devices. However, aggregating large volumes of sensitive patient information at a central server poses several privacy, security, and regulatory hurdles. This paper proposes a decentralized Federated Learning (FL) architecture that can accommodate multi-device IoT-based medical data integration without sharing sensitive patient information between healthcare institutions. Each institution maintains their unique model, which can be used for training on their premises, and then the models are integrated into a single federated model for collective use. A novel framework is proposed, which can accommodate collaborative training between several nodes of healthcare institutions, along with local processing of multimodal data obtained from IoT medical devices. The multimodal deep neural network architecture has been used to combine different types of input, including medical images and structured demographic or clinical attributes, through parallel feature extraction branches followed by feature-level fusion. Each healthcare institution has a unique model, which is trained on their respective data, and then the model parameters are sent to a central server for aggregation through the Federated Averaging (FedAvg) algorithm. Cross-validation of the models has also been done to check the robustness of the models, which involves training on several nodes and then validation on an external dataset. The results obtained through the federated multimodal framework indicate high classification accuracy, thus ensuring the viability of a secure, scalable, and privacy-preserving collaborative intelligence system for IoT-based healthcare environments.},
month = {March},
doi = {https://doi.org/10.64388/IREV9I9-1715262}
}