AUTHOR=Srour Tarek , El-Rifaie Ali M. , El-Bendary Mohsen A. M. , Eltokhy Mostafa , Abouelazm Atef E. , Neji Bilel 

TITLE=Multimedia privacy protection: an N-round cascaded cryptosystem based on merged multi-chaotic maps under various image attacks

JOURNAL=Frontiers in Computer Science

VOLUME=Volume 7 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/computer-science/articles/10.3389/fcomp.2025.1551166

DOI=10.3389/fcomp.2025.1551166

ISSN=2624-9898

ABSTRACT=Due to the widespread use and variety of wireless networks and their applications in open environments, security techniques need to be more robust, reliable, and flexible, incorporating multi-stage and interfered secret key (Skey) capabilities. This research article presents an efficient cryptographic approach through the construction of a cascaded series of processes. It proposes a multi-number of encrypting processes (N-round chaos-based) cryptosystem containing N secret keys based on data classification and the environment. The article also outlines the simple criteria for determining the number of rounds (N-round). Several chaos-based encryption techniques have been utilized to construct the proposed algorithm, and various scenarios of the N-round mechanism over different classified images have been presented. In the majority of relevant published studies, two critical issues have been overlooked: the selection criteria for required and suitable security levels and the rules and conditions for robust security keys. These essential issues have been discussed in this research article. Various grayscale images are used in the computer simulation experiments conducted to demonstrate the robustness, reliability, and applicability of the proposed cryptographic algorithm. The experiments consider the presence of various attacks, such as speckle, salt-and-pepper, and Gaussian noise. Based on the results from the standard database (Hlevkin, USA CANVAS), the proposed algorithm is robust and suitable for applications in noise/attack environments. As clarified by the results, the similarity of extracted images is 100% to the original plaintext in the absence of attacks, while in the presence of noise and attacks, the similarity remains above 99%. The quality of the decrypted images is superior to that of the majority of existing cryptosystems.