Improve steganography system using agents software based on statistical and classification technique

Abstract

In digital communications, information security is a paramount necessity. In the hiding algorithm, there are three basic parameters: security, capacity, and imperceptibility. Therefore, there are many ways to design the steganography algorithm, such as least significant bit (LSB), discrete wave transformation (DWT), and discrete cosine transform (DCT). The aim of this paper is to improve agent software design based on a steganography system. It proposed an agent system based on a support vector machine (SVM) classifier to hide a secret message in a certain cover image. The common dataset for steganography uses 80% training and 20% testing to get accurate results. Developing an agent system depends on six statistical parameters such as energy, standard deviation, histogram, variance, mean, and entropy. This resulted in features classified by the SVM classifier to predict the best cover image to be nominated for embedding. Worthy results were obtained in terms of imperceptibility, attack, and cover image prediction by statistical issues.