Majlesi Journal of Electrical Engineering IOT-MDEDTL: IoT Malware Detection based on Ensemble Deep Transfer Learning Volume 16 (2022) Issue 3, September 2022 2024 02 11 IOT-MDEDTL: IoT Malware Detection based on Ensemble Deep Transfer Learning 10.30486/mjee.2022.696506 EN Qasim Kadhim English Language Department, Al-Mustaqbal University College, Babylon, Iraq Ahmed Qassem Ali Sharhan Al-Sudani Al-Manara College For Medical Sciences, Maysan, Iraq Inas Amjed Almani Department of Computer Technology Engineering, Al-Hadba University College, Iraq Tawfeeq Alghazali College of Media, Department of Journalism, The Islamic University in Najaf, Najaf, Iraq Hasan Khalid Dabis College of Islamic Science, Ahl Al Bayt University, Kerbala, Iraq Atheer Taha Mohammed The University of Mashreq, Iraq Saad Ghazi Talib Law Department, Al-Mustaqbal University College, Babylon, Iraq Rawnaq Adnan Mahmood Medical device engineering, Ashur University College, Baghdad, Iraq Zahraa Tariq Sahi Department of Dentistry, Al-Zahrawi University College, Karbala, Iraq Yaqeen Mezaal Al-Esraa University College, Baghdad, Iraq Journal Article 2024 02 11 The internet of Things (IoT) is a promising expansion of the traditional Internet, which provides the foundation for millions of devices to interact with each other. IoT enables these smart devices, such as home appliances, different types of vehicles, sensor controllers, and security cameras, to share information, and this has been successfully done to enhance the quality of user experience. IoT-based mediums in day-to-day life are, in fact, minuscule computational resources, which are adjusted to be thoroughly domain-specific. As a result, monitoring and detecting various attacks on these devices becomes feasible. As the statistics prove, in the Mirai and Brickerbot botnets, Distributed Denial-of-Service (DDoS) attacks have become increasingly ubiquitous. To ameliorate this, in this paper, we propose a novel approach for detecting IoT malware from the preprocessed binary data using transfer learning. Our method comprises two feature extractors, named ResNet101 and VGG16, which learn to classify input data as malicious and non-malicious. The input data is built from preprocessing and converting the binary format of data into gray-scale images. The feature maps obtained from these two models are fused together to further be classified. Extensive experiments exhibit the efficiency of the proposed approach in a well-known dataset, achieving the accuracy, precision, and recall of 96.31%, 95.31%, and 94.80%, respectively.