DoS and DDoS Attack Detection in IoT Infrastructure using Xception Model with Explainability
Year: 2025
Type: conference-paper
Abstract: The denial of service (DoS) and distributed denial of service (DDoS) attacks are considered the most frequent attacks targeting the Internet of Things (IoT) network infrastructure globally. The current approaches for detecting DoS and DDoS attacks mainly use intrusion detection systems, traffic monitoring, and firewalls. However, complex DoS and DDoS attacks can bypass these detection mechanisms. Thus, this paper proposes utilizing convolutional neural network-based transfer learning to detect DoS and DDoS attacks from converted network traffic data into images. We employed the Xception model with fine-tuning, and we achievedan average of 91% accuracy in detecting eleven different types of DoS and DDoS attacks, which is higher than the current state-of-the-art by 5% targeting the same task.
Authors: Nelly Elsayed, Zag ElSayed, Ahmed Abdelgawad
Institutions: University of Cincinnati, Central Michigan University
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4410398247