Development of a Subsystem for Steganalysis of Digital Images Based on a Convolutional Neural Network to Detect and Prevent Attacks Using Hidden Steganographic Channels

Download article in PDF format

Authors: Yandashevskaya E. A.

Annotation: This article presents a way to implement the subsystem for steganalysis of digital images circulating in the information system. This subsystem expands the functionality of existing intrusion detection / prevention systems in terms of detecting covert channels used in computer attacks. In the presented solution, a parametric model of a convolutional neural network is proposed and implemented to detect a payload in digital images, performed by a number of steg-nesting algorithms recognized in real attacks. A software implementation of a modular generator of a training sample (dataset) that supports these algorithms has been developed. An experimental assessment of the accuracy has been carried out.

Keywords: information protection, hidden steganographic channels, convolutional neural network, intrusion detection and prevention systems, digital images

Editorial office address

Executive Secretary of the Editor’s Office

 Editor’s Office: 40 Lenina Prospect, Tomsk, 634050, Russia

  Phone / Fax: + 7 (3822) 701-582

  journal@tusur.ru

 

Viktor N. Maslennikov

Executive Secretary of the Editor’s Office

 Editor’s Office: 40 Lenina Prospect, Tomsk, 634050, Russia

  Phone / Fax: + 7 (3822) 51-21-21 / 51-43-02

  vnmas@tusur.ru

Subscription for updates