An image encryption method based on a dynamic Hopfield–Voronoi coupled map lattice
Xin Xie, Zhiyu Chen, Hao Ning, Yu Zhou, Zhengyu Li, Kun Zhang · Journal of King Saud University - Computer and Information Sciences · 2026
Chaos-based image encryption requires digital chaotic sources that retain complex and weakly correlated dynamics under finite-precision computation. This study presents a context-adaptive encryption framework based on a dynamic Hopfield–Voronoi coupled map lattice (DHV-CML). Sparse Hopfield-inspired neural units use regional feedback to regulate lattice evolution and dynamically reorganize spatial interactions. The resulting trajectory generates key material for adaptive blockwise Hilbert scanning, global pixel permutation, and reversible bidirectional feedback diffusion. Dynamical analyses show that DHV-CML mitigates periodic degradation, maintains broad chaotic coverage and high Kolmogorov–Sinai entropy density, and satisfies the reported NIST STS criteria. Encryption experiments demonstrate low statistical leakage, strong key sensitivity, effective differential diffusion, and resistance to noise and cropping attacks. Complexity and timing analyses further quantify the computational overhead and practical throughput. These results support DHV-CML as an effective adaptive chaotic source for secure image encryption.