Multi Resolution Decomposition For A Passive Steganalysis Based On a Multi Agent System
Abla Bouguerne, B. Hayet, Farida Merouani, C. Nouha Kobsi · 2012
Abstract- This article proposes a methodology for steganalysis using feature selection, higher order based on the wavelet decomposed image to bring out the best properties and detect the presence of any hidden message in the image, we had specified the JPEG format because it is the most commonly used for hiding information. As based on this principle, a model of universal steganalysis is proposed, using the first order statistics and the entropy to detect deviations in the image containing a hidden message. The proposed method of steganalysis based on Multi Agent System (SMA) is capable, with the presence of the covermedium, to reveal whether an image contains a hidden message or not, each agent has to solve a specific task.