Adaptive batch color image encryption algorithm based on image hash-driven and multi-round chaotic perturbation

Abstract

Existing chaotic image encryption algorithms often suffer from fixed initial parameters and insufficient adaptability to image content, limiting their resistance to specific attacks and efficiency in batch processing. Addressing these deficiencies, this paper proposes an adaptive batch color image encryption algorithm based on image hash-driven techniques and multi-round chaotic perturbation. By utilizing the SHA-256 hash of the plain image as a dynamic seed, the algorithm integrates a novel Sine-Tent-Cosine (STC) mapping and an improved Chebyshev-Tent (CT) mapping to realize content-sensitive dynamic parameter adjustment. To enforce deep encryption, a hierarchical perturbation chain is constructed: the Rubik’s cube layer executes 3D spatial scrambling via chaos-driven multi-layer rotations to disrupt pixel positions; the triple diffusion mechanism employs bitwise XOR, global cyclic shifts, and mirror reversal operations to trigger a robust avalanche effect; and the Fisher-Yates shuffling further eliminates statistical correlations through chaos-guided global permutation. Empirical results substantiate the advantage of the proposed approach, with the encrypted images attaining an information entropy of 7.9992 and successfully passing the NIST randomness tests. The algorithm maintains strong robustness against differential attacks, with NPCR and UACI values recorded at 99.6098% and 33.4606%, respectively. Efficiency evaluation shows an encryption time of 0.0119s for a single


$$256 \times 256$$


256
×
256




image and 0.1091s for a batch of 10 images. Consequently, this study offers a dependable solution to support the secure transmission of large-scale image data.