Development of cryptographic techniques for images using discrete rajan transform

Abstract

The objective of image cryptography is to secure the image over the public networks. In emerging technology, hacking of data has become a critical issue due to which safe encryption scheme is becoming a foremost need for everybody. To accomplish these requirements, various traditional encryption algorithms are developed. However, due to some inbuilt properties of images such as huge data capacity, high redundancy, the high correlation among pixels and low entropy, these algorithms are inappropriate for image encryption. Transform domain encryption methods comprise mathematical transforms for fast computation. The fractional orders in fractional wavelet transform, vector parameters in fractional Fourier transform serve as relatively small cipher keys. The transforms such as Discrete Wavelet Transform, Discrete Cosine Transform are used to extract the coefficients for encryption. The transforms are not the part of the encryption, but their parameters are used as the keys. Hence the image or information is not communicated securely. However, all of the above techniques have failed to strengthen their mathematical representations and algorithms to generate standard keys. The issues discussed above, with the possibility of further avenues for improving image security performance, served as motivation and basis for the proposed work. The aim of this thesis is to determine the security of images using transform in terms of statistical, differential attack analysis, and reconstructed image quality. The thesis presents three modules of security algorithms and their performance analyses. In the first methodology, image cryptography is developed based on Discrete Rajan Transform (DRT) and cyclic shift invariance. The pixels are shifted cyclically, and DRT is applied to generate transformed pixels along with encryption keys. newline