Data Security Challenges in Deep Neural Network for Healthcare IoT Systems

Edmond S. L. Ho*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

Abstract

With the advancement of IoT technology, more and more healthcare applications were developed in recent years. In addition to the traditional sensor-based systems, image-based healthcare IoT systems become more popular since no specialized sensors are required. Combining with Deep Neural Network (DNN) based automated diagnosis and decision-making systems, it is possible to provide users with 24/7 health monitoring in real life. However, the high computational cost for training DNNs can be a hurdle for developing such kind of powerful systems. While cloud computing can be a feasible solution, uploading training data for the DNN models to the cloud may lead to data security issues. In this chapter, we will review some image-based healthcare IoT systems and discuss some potential risks on data security when training the DNN models on the cloud.
Original languageEnglish
Title of host publicationSecurity and privacy - preserving for IoT and 5G networks: Techniques, challenges, and new directions
Place of PublicationCham
PublisherSpringer
Publication statusAccepted/In press - 26 Apr 2021

Fingerprint

Dive into the research topics of 'Data Security Challenges in Deep Neural Network for Healthcare IoT Systems'. Together they form a unique fingerprint.

Cite this