Image encryption algorithm based on COA and hyperchaotic Lorenz system

Xiaofu Qin,Yong Zhang

DOI: https://doi.org/10.1007/s11071-024-09632-6

IF: 5.741

2024-05-08

Nonlinear Dynamics

Abstract:Aiming at the problems of weak security and susceptibility to violent cracking in traditional image encryption algorithms, this paper presents a novel image encryption algorithm based on the Coyote Optimization Algorithm (COA) and the hyperchaotic Lorenz system. The hyperchaotic Lorenz system exhibits sensitivity to both initial conditions and parameters, and complex dynamics behavior. These characteristics pose a challenge for attackers attempting to extract essential information from the image through analysis, thereby enhancing the algorithm's resistance to cracking. Nevertheless, the hyperchaotic Lorenz system is susceptible to the initial values of state variables, and its initial parameters can be easily deciphered. Consequently, this study suggests employing the COA to optimize the sequence generated by the chaotic system. This is done to increase the randomness and complexity of the key, making it more challenging to crack. Given that COA gets trapped in local optima when dealing with high-dimensional problems, this paper proposes the Lévy-flight Coyote Optimization Algorithm (LCOA). By implementing the LCOA, which involves larger step sizes and faster jumps, the algorithm is expected to achieve global optimality during the optimization search process. The experimental results demonstrate that the optimization results of the generated sequence of the hyperchaotic Lorenz system by LCOA can eliminate patterns and regularities of pixel points in the image generated effectively. This leads to a significant enhancement in the resistance of the image encryption algorithm to differential attacks, resulting in a significant enhancement in image encryption performance.

engineering, mechanical,mechanics