High Spatial Capacity Steganographic Algorithms based on Predictive Coding

Abstract

Information hiding refers to embedding additional digital data in cover objects–e.g. audio,images or video signals – by modifying the cover objects. These techniques are used in a variety of application domains like media forensics, content authentication, copyright protection and covert communications. The main objective of information hiding techniques is to minimize the effect that the embedding process has on the carrier object. Embedding processes introduce distortion to cover images and change the appearance as well as primary statistics of images. In this dissertation, information hiding algorithms using both colored and grey images have been proposed with three basic goals: improving peak signal to noise ratio, preserving statistical properties and predicting a distortion level. The prediction errors created by the differential and brightness matrix for JPEG quantization used as improving spatial hiding information. The techniques used are lossless predictive coding and adaptive predictive coding. Adaptive predictive coding is an improved hybrid technique concatenating both prediction error expansion method and difference expansion. It is evident from the results that techniques introduced by proposed algorithm are efficient, functional methods for maximum hiding capacity and improved peak signal to noise ratio of the image. Minimization of distortion has been experimentally validated for the improved algorithms.