A Prediction Model Based Approach to Open Space Steganography Detection in HTML Webpages



Coenen, FP ORCID: 0000-0003-1026-6649, Sedeeq, I and Lisitsa,
(2017) A Prediction Model Based Approach to Open Space Steganography Detection in HTML Webpages. .

[img] Text
iwdw2017.pdf - Author Accepted Manuscript

Download (293kB)

Abstract

A mechanism for detecting Open Space Steganography (OSS) is described founded on the observation that the length of white space segments increases in the presence of OSS. The frequency of white space segments of different length is conceptualized in terms of an n-dimensional feature. This feature space is used to encode webpages (labelled as OSS or not OSS) so that each page is represented in terms of a feature vector. This representation was used to train a classifier which can subsequently be used to detect the presence, or otherwise, of OSS in unseen webpages. The proposed approach is evaluated using a number of different classifiers and with and without feature selection. Its operation is also compared with two existing OSS detection approaches. From the evaluation a best accuracy of 96.7% was obtained. The evaluation also demonstrated that the proposed method outperforms the two alternative techniques by a significant margin.

Item Type: Conference or Workshop Item (Unspecified)
Uncontrolled Keywords: Open space, Steganography, Classification
Depositing User: Symplectic Admin
Date Deposited: 12 Jun 2017 07:04
Last Modified: 19 Jan 2023 07:03
DOI: 10.1007/978-3-319-64185-0_18
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3007907