A new fast high dimensional and memristive hyperchaotic multiple image encryption method and its FPGA implementation
Wassim Alexan, Youstina Megalli
Abstract
Abstract This article presents a high-throughput image encryption algorithm implemented on an Artix-7 FPGA, which offers parallel processing, low latency, and real-time performance with a throughput of 0.8 Gbps, making it ideal for secure applications such as satellite imaging, surveillance, and medical diagnostics. The proposed method employs three chaotic systems: a 10D hyperchaotic system, an 8D hyperchaotic system, and a memristive coupled neural network (MCNN), to achieve a multi-layered encryption process that enhances confusion, diffusion, and key space complexity. The total key space of the algorithm is approximately $$2^{3454}$$ , providing an extremely large search space that ensures robust protection against brute-force attacks. Performance results demonstrate excellent security properties, including high entropy values ( $$\approx 7.999$$ ), low pixel correlation coefficients (average $$\le 0.003$$ ), and resistance to differential attacks. The FPGA-based implementation significantly outperforms software-only solutions in both speed and efficiency, validating its suitability for real-time, high-security image encryption scenarios.