Blind steganalysis using fractal features

A novel approach for detecting Steganographic images with blind steganalysis using fractalfeatures has been proposed in this thesis. Two overarching methods were used to constructthe feature vector; first, using a variation of the Differential Box Counting algorithm forlacunarity estimation to extract the fractal features; and then, using dynamic time warpingfor similarity measures as the basis for further deriving other features.The research design enabled the proposition of four major approaches that were based oniterative experiments that aided in further improving and extending upon the previousoutcomes.This research has thus made three major contributions to the body of knowledge by thefollowing:1. Proposing of a novel approach for constructing the feature vector based on fractalfeatures for blind steganalysis.2. Ability to perform significant feature reduction by using the proposed fractal fea-tures, which is also applicable in areas other than steganalysis.3. Discovery of an improved blind steganalysis approach for known Cover images.

Grantor dc:publisher

Edith Cowan University, Research Online, Perth, Western Australia

Year dc:date

2016

Author dc:creator

• Ibrahim, Ahmed

• Steganography
• Steganalysis
• Blind Steganalysis
• Fractal Features
• Lacunarily
• Dynamic Time Warping
• Support Vector Machines
• Outguess
• F5
• StegHide
• Computer Sciences

Repository record dc:identifier

https://ro.ecu.edu.au/theses/2198

OAI identifier oai:identifier

oai:ro.ecu.edu.au:theses-3200