A novel symmetric cryptographic algorithm to enhance the security of multimedia (image) information in IoTs

Abstract

The rapid expansion of the Internet of Things (IoT) has exposed critical weaknesses in existing security frameworks, mainly in widely deployed, resource-constrained devices. Studies indicate that nearly 70% of IoT devices are vulnerable to severe security threats, making multimedia data, especially images, highly susceptible to interception, tampering, or unauthorized access to secret images. Securing images in IoT environments is challenging due to their high redundancy and strong pixel correlations. Existing image encryption algorithms, both traditional and modern, suffer from static architectures, predictable plaintext-ciphertext correlations, fixed key-expansion patterns, imbalanced frequency distributions, and reduced entropy, which weaken their resilience against modern statistical, linear, and differential cryptanalytic attacks. To overcome these limitations, we propose a symmetric-based hybrid multi-layer hyper-chaotic image encryption scheme, a dynamic key mixer, adaptive rule-based permutation, and an independent nonlinear key scheduling module. These mechanisms transform the original images into a strongly ciphered image representation with enhanced confusion, diffusion, and cryptographic robustness, ensuring resistance to various attacks. Security analysis of our proposed schemes demonstrated superior performance, achieving a near-ideal global entropy of 7.99957, a correlation coefficient of − 0.000123, and a pixel change rate (NPCR) of 99.75563%. These metrics indicate state-of-the-art resistance against statistical, known-plaintext, and chosen-plaintext attacks. Extensive benchmarking against conventional standards, including DES, Triple DES, and AES, as well as contemporary chaotic-based frameworks, reveals the unprecedented achievements of the proposed method in terms of key space and complexity. The proposed method ensures high unpredictability and a significant security margin by expanding the bit-level complexity to 2247, 4295, and 8391 bits for the 256-bit, 512-bit, and 1024-bit configurations, respectively. The system achieves formidable security margins of 2 2247 , 2 4295 , and 2 8391 . These results confirm that the proposed algorithm provides robust randomness, high security, and effective resistance to differential and statistical attacks.