A Hybrid Image Encryption Scheme Based on Cellular Automata and RSA

Abstract Technological innovations have made it possible to share digital images via various image-sharing services, enabling easy, widespread distribution. However, conventional encryption methods are often inefficient for securing image information because they do not effectively handle image size, high redundancy, and strong inter-pixel correlation characteristics. In this paper, we introduce a hybrid image encryption method that combines RSA-based key encapsulation with the pseudo-random keystream generation capability of cellular automata. Image security is achieved through three components: first, RSA-OAEP is used to encapsulate the SHA-256 hash-derived encryption parameter instead of encrypting the image pixels directly; second, cellular automata with distinct rules are used for different color channels to introduce statistical randomness into the image data; and third, encryption and scrambling operations are applied to the image pixels to generate noise-like cipher images with no visually recognizable original patterns. The proposed method produces noise-like cipher images while preserving lossless reconstruction quality after decryption. The scheme achieves near-uniform histograms, entropy close to 8, NPCR around 99.60%, UACI close to 33.46%, and linear complexity $$O(MN)$$ , with average encryption times of $$0.065596$$ , $$0.305842$$ , and $$1.144842$$ s for $$256\times 256$$ , $$512\times 512$$ , and $$1024\times 1024$$ images, respectively.