ENHANCEMENT OF THE SECURITY OF DIGITAL IMAGES USING SOME MODULI SETS

Abstract

Globally digital images have found usage in diverse fields. These images may contain confidential information and need to be protected when stored in memory or transmitted over networks. Many techniques have been proposed over the years to deal with these security issues. The Arnold's transform is an image scrambling technique that have been extensively refined and used for image encryption. However, most of the reported works are applicable to only square images and also periodic. In this thesis, we propose two new security enhancement cryptosystems for digital images. The schemes employ two methods: residue number system to decimal (RID) encoding and decoding and a modified Arnold's transform algorithm or bitwise-XOR operation. The encryption process of our first scheme (Scheme I) resizes a plain-image, encodes it using the moduli set [Z" - 1, 2n, 2n + 1} and finally scrambles it using a modified Arnold's transform. In the decryption process, the cipher-image is scrambled, decoded and then resized to recover the plain-image. The second scheme (Scheme II) resizes a plain-image, scrambles it with a bitwise-XOR operator and then encodes it with moduli set {2n + 2, 2n + 1, 2n}. The decryption process decodes, scrambles and resizes a cipher-image into its plain equivalent. The proposed schemes were simulated using MATLAB. Preliminary results show that the schemes can effectively encrypt and decrypt images without lost of inherent information. The schemes also offer firm resistance to statistical attacks such as histogram, bruteforce, correlation coefficient and key sensitivity.