Symmetric cryptosystems based on chaotic maps for images devoid of size constraints

Abstract

The enormous growth in communication networks and communication devices has created great potential for multimedia generation and transmission, especially in terms of images. Digital images are vulnerable to security threats, with security concerns looming large. Consequently, there is a need for security measures to be put in place, and a variety of encryption schemes have been proposed to address this challenge. Schemes formulated for textual data are not suited to images, owing to various inherent features as high pixel frequency, volume of data, and close correlation among pixels. As a solution, chaos-based cryptosystems have emerged as a field in cryptography and found to deal with images efficiently. A well-defined chaos based image cryptosystem is a combination of confusion and diffusion processes. In the confusion process, the pixel positions are changed and in the newlinediffusion process, pixel values are changed. In this thesis, appropriate chaotic maps are identified for use in the confusion and the diffusion phases of image encryption. Techniques based on these chaotic maps are combined together to form desirable image cryptosystems. The objective of this work is to develop and implement a cryptosystem for images having unequal dimensions. The Arnold transform is the simplest and frequently-used confusion technique. The basis matrices of the Arnold transform are combined together with their mirrors to produce new enhanced transformation matrices. The enhanced Arnold transform (EAT) matrices can be used in the confusion phase of any image cryptosystem to mystify the relationship between the plain image and the scrambled image. newline newline