Mixed-multiple image encryption algorithm using RSA cryptosystem with fractional discrete cosine transform and 2D-Arnold Transform

This paper presents a novel multiple-image encryption (MIE) technique based on mixed image elements associated with RSA cryptosystem, fractional discrete cosine transform (FrDCT) and 2D-Arnold transform (AT). Firstly, we combine k-original images into a big image using knowledge of matrix theory. The big image is extracted into three components, i.e., red (R), green (G) and blue (B). The RSA cryptographic scheme is applied to each component R, G and B individually. Secondly, the FrDCT is used on each component with individual fractions. Thirdly, the 2D-AT is applied on three components to enhance the security as well as the key space. Finally, three components are concatenated to generate a final encrypted image. The resulting cipher image is a single-channel real-valued image that is easy to display, store and transmit over an unsecured network. The proposed algorithm offers multi-layered security in frequency, time and coordinate domains. Security of the presented MIE technique depends on both secrete keys and their proper arrangement. Simulation analysis and their results support the robustness and appropriateness of the introduced encryption technique. Sensitivity analysis confirms that the introduced technique is extremely sensitive to its private keys and their arrangement. The effectiveness and viability of our proposed cryptosystem are confirmed by statistical analysis, including MSE, PSNR, SSIM, entropy analysis, noise attack, correlation coefficient, occlusion attack analysis, histogram analysis, UACI, NPCR, classic types of attacks and comparison analysis. The proposed algorithm is proved to be very effective and secure for real-world image encryption, which is supported by the experimental results and algorithm analysis.