A Robust Image Cryptosystem Incorporating Bit-Level Swapping, Dynamic Multi-Map Substitution, And Iterative Block Diffusion

Abstract

The rapid expansion of digital communication has necessitated the development of advanced cryptographic frameworks capable of securing visual data against sophisticated cryptanalytic threats. This paper introduces a robust, tripartite image cryptosystem designed to address the inherent redundancies and high spatial correlations found in digital images. Unlike traditional pixel-shuffling methods, the proposed architecture operates at the bit-stream level through three integrated stages: Bit-level Recombination and Permutation, Dynamic Multi-Map Chaotic Substitution, and a Multi-round Iterative Diffusion Layer. The permutation phase utilizes a symmetrical interchanging process to effectively neutralize spatial redundancies. A key innovation in the substitution layer is the implementation of a key-driven dynamic selection mechanism that switches between four distinct chaotic systems—Tent, Lorenz, Logistic, and Henon maps—ensuring session-specific unpredictability and resistance to modeling attacks. Security is further fortified by a block-cipher-based iterative diffusion layer incorporating XNOR-based logical masking and feedback loops to propagate minor input variations throughout the entire ciphertext.
Empirical validation via MATLAB simulations demonstrates that the proposed scheme achieves near-ideal performance benchmarks, including an information entropy of approximately 7.9993 and a perfectly uniform histogram distribution. Furthermore, the system exhibits exceptional resistance to differential attacks, yielding a high Number of Pixels Change Rate (NPCR) and Unified Average Changing Intensity (UACI) scores that outperform several contemporary methods. Additionally, the cryptosystem demonstrates high computational efficiency, with a linear-time complexity of O(M×N), making it an ideal candidate for real-time secure communication and high-definition video encryption. Comparative analyses confirm that the bit-level granularity and iterative diffusion layers successfully eliminate inter-pixel correlations, providing a secure and computationally efficient solution for real-time image protection in modern communication environments.