Efficient and secure image compression and encryption based on compressive sensing and four-dimensional hyperchaotic system

Abstract

With the growing awareness of privacy protection, the security of image data during network transmission has become a focal point of concern. Reducing computational complexity and improving transmission efficiency while meeting high-security requirements has become a primary focus of current research. To address this, we propose a novel image privacy protection scheme that combines compressive sensing with chaotic encryption, aiming to ensure image privacy and security while enhancing transmission and storage efficiency. We employ compressive sensing technology to achieve efficient compression of image data. Unlike traditional compression and encryption schemes, the proposed method does not require explicit sparsification preprocessing, thereby avoiding the complex operations introduced by signal transformations and simplifying the signal recovery process. To enhance encryption security, a four-dimensional hyperchaotic system with stronger chaotic properties is designed to generate highly random and unpredictable key streams, ensuring the security of the encrypted data. Furthermore, this paper combines disjoint Latin squares with fractal generation strategies to design a new fractal index matrix, based on which a novel image permutation scheme is proposed. This scheme effectively eliminates the linear relationships and correlations between adjacent pixels, achieving global pixel permutation. Coupled with the proposed dual-channel bidirectional diffusion structure, the algorithm effectively diffuses pixel information across the entire image, increasing the complexity and unpredictability of the image encryption process. Experimental results indicate that the proposed algorithm exhibits excellent performance in terms of compression efficiency, encryption effectiveness, and resistance to attacks, providing a highly efficient and reliable solution in the field of image compression and encryption.