A flexible image cipher based on orthogonal arrays (Q2126221): Difference between revisions

The paper is written in excellent chronological order by first introducing several research papers from the bibliography on image encryption and chaotic systems. The authors point out the paper of Rao who introduced orthogonal arrays in the 1940s. I feel some of the work of Turing also plays a part in breaking encryptions. They also point out that cryptosystems are defined on finite sets whereby chaotic systems have meaning on the real number field. The paper moves forward on how to enhance higher levels of security of the cryptosystem. A previous historical technique uses Latin Squares. A Latin Square is an \(n\times n\) array comprised of a set of symbols. Each row and column of the array contains each symbol the same number of times. The authors introduce a combinatorial configuration of an orthogonal array to image encryption. They define an orthogonal array of size \(n\), with \(q\) constraints, \(s\) levels and strength \(t\), denoted \(OA(n,q,s,t)\). They nicely include an example for a transpose of an array to be \(OA(9,4,3,2)\) having \(9\) rows, \(4\) columns, \(s=\{1,2,3\}\) symbols, which generates \(9\) distinct ordered pairs. The distinct \(9\) ordered pairs appear exactly once in any two columns. Then any two columns of the \(OA(9,4,3,2)\) can generate a 2D: \(1\) to \(1\) mapping and the authors exhibit one such mapping in the paper. This sets the stage to generalize the methodology and software written contained in the paper. The methodology and software programs contained in the paper were developed by Ming Xu and the conceptualization by Zihong Tian. The paper includes five algorithms. The paper then goes into simulation results and efficiency analysis. The authors include six well-known images. The first image Lena contains a variety of gradations and textures that proved to be a challenge in processing algorithms. They include the images of histograms of the correlation distribution of plaintext, images of cipher images, and the correlation distribution of cipher images. The results then include a detailed analysis of the entropy analysis together with a detailed diffusion analysis together with a known chosen plaintext attack analysis. The algorithms include several keys which can be public or secret depending on how much security is needed. I find the paper a joy to read and review very deep notions of cyber security.