Deep-Learning-Empowered Digital Forensics for Edge Consumer Electronics in 5G HetNets

Feng Ding, Guopu Zhu, Mamoun Alazab, Xiangjun Li, Keping Yu

Research output: Contribution to specialist publicationArticle

Abstract

The upcoming 5G heterogeneous networks (HetNets) have attracted much attention worldwide. Large amounts of high velocity data can be transported by using the bandwidth spectrum of HetNets, yielding both great benefits and several concerning issues. In particular, great harm to our community could occur if the main visual information channels, such as images and videos, are maliciously attacked and uploaded to the internet, where they can be spread quickly. Therefore, we propose a novel framework as a digital forensics tool to protect end users. It is built based on deep learning and can realize the detection of attacks via classification. Compared with the conventional methods and justified by our experiments, the data collection efficiency, robustness, and detection performance of the proposed model are all refined. In addition, assisted by 5G HetNets, our proposed framework makes it possible to provide high-quality real-time forensics services on edge consumer devices (ECE) such as cell phones and laptops, which brings colossal practical value. Some discussions are also carried out to outline potential future threats.

Original languageEnglish
Specialist publicationIEEE Consumer Electronics Magazine
DOIs
Publication statusAccepted/In press - 2020

Keywords

  • 5G mobile communication
  • Consumer electronics
  • Deep learning
  • deep learning
  • Detectors
  • digital forensics
  • Digital forensics
  • edge consumer electronics
  • Forensics
  • Information security
  • Tools

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Hardware and Architecture
  • Computer Science Applications
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Deep-Learning-Empowered Digital Forensics for Edge Consumer Electronics in 5G HetNets'. Together they form a unique fingerprint.

Cite this