Abstract
Issues such as inefficient encryption architectures, nonstandard formats of image datasets, weak randomness of chaos-based Pseudorandom Number Generators (PRNGs), omitted S-boxes, and unconvincing security metrics leading to increased computational time and inadequate security level of chaos and Deoxyribonucleic Acid- (DNA-) based image encryption schemes need careful examination towards the development of more stable encryption schemes in terms of efficiency and reasonable security. A new taxonomy of image encryption based on chaotic systems, hyperchaotic systems, and DNA is propounded to assess the impact of these issues on the performance and security metrics. The primary emphasis of this research is to study various recent encryption architectures centered on a variety of confusion and diffusion methods. It is aimed at assessing the performance and security of various ciphers using a cipher rating criterion that categorizes ciphers into different classes. The parameters that are included in the rating criteria are information entropy, chi-squared goodness of fit test for histogram uniformity analysis, encryption efficiency, key space, differential attacks (Number of Pixels Change Rate and Universal Average Changing Intensity), key sensitivity analysis, encryption time, randomness tests such as NIST-R (a statistical suite for validating the randomness designed by the National Institute of Standards and Technology), correlation coefficient analysis, contrast analysis, energy analysis, homogeneity analysis, Mean Absolute Error, peak signal-to-noise ratio, and robustness to noise and occlusion attacks.