Abstract
Chaotic systems have attracted considerable attention in image encryption due to their complex nonlinear dynamics, sensitivity to initial conditions, and ability to generate pseudorandom sequences. Motivated by the need to balance chaotic complexity, controllability, statistical randomness, and computational efficiency, this study introduces a novel three-dimensional discrete chaotic map derived from a dark matter–dark energy interaction model using one-step Euler discretization. The objective is to characterize its nonlinear dynamics and exploit its mathematical structure for image encryption rather than investigate the cosmological properties of the original model. The proposed map is analyzed through equilibrium stability, phase portraits, Lyapunov exponents, bifurcation diagrams, complexity, and spectral entropy, revealing complex chaotic behavior and positive Lyapunov exponents. A feedback control scheme is also developed to stabilize the system toward desired equilibrium states. The generated chaotic sequences are integrated into a layered image encryption framework based on a pre-shared key (PSK), a random initialization vector (IV), and SHA-256-based key derivation. The resulting key-dependent sequences drive bit-level permutation and bidirectional dynamic diffusion using modular arithmetic, bitwise XOR, and chaining operations. The cryptosystem is evaluated through histogram, correlation, entropy, key-space, key-sensitivity, NPCR, UACI, NIST SP 800-22, and robustness analyses against noise and cropping attacks. The results demonstrate favorable statistical and differential properties, high key sensitivity, robustness to transmission degradations, and satisfactory computational efficiency, highlighting the potential of the proposed discrete chaotic framework for secure image encryption.