Deep Learning IoT Malware Detection Model for IoMT Edge Devices

Abstract

Internet of Things (IoT) is defined as the massive collection of physical devices being connected to the Internet. IoT has a positive impact in multiple fields, such as health, agriculture, and power management sectors by advancing them to new technical horizons. However, such advanced technologies introduce security challenges that can negatively affect IoT applications and possibly threaten their existence. In the health sector, for instance, Internet of medical things (IoMT) devices are used to perform tasks such as remote patient monitoring and to gather biometric information. Also, these devices are used as a base for several healthcare procedures such as prescribing medication. Several security breaches can occur to IoMT devices that may expose human privacy and security since the data collected and processed is very sensitive. In this thesis, we provide a light-weight malware detection deep learning model. The model is deployed on IoMT edge devices that can detect IoT specific malware. The proposed models utilize gray-scale images produced by the binary of malware files to classify malware from goodwares. The achieved results were promising in terms of malware classification accuracy, which might help prevent malware and secure the dedicated systems for IoMT devices and applications.