Images Within Images? A Multi-image Paradigm with Novel Key-Value Graph Oriented Steganography

Abstract

Steganographic methods have been in the limelight of research and development for concealing secret data within a cover media without being noticed through general visualization. The Least Significant Bits (LSBs) of 8bit color code for the RGB image arises the possibility of replacing the last two bits with the bits of the encrypted message. Several procedures have been developed to hide an image within another image however in most cases the payload image has to be within the accommodatable range of the cover image and very little literature have shown methods to hide multiple images within multiple images. This paper presents a novel approach to split the image into JSON styled dictionary of key-value pairs and using a metadata graph to locate different parts and positions of the payload images in the entire cluster of cover images. The model could be easily used in the real world scenario for privately sharing secret data over public communication channels without being noticed.