A data hiding scheme in encrypted images using non-degenerate chaotic system and hybrid coding A data hiding scheme in encrypted images

• Published in: Nonlinear dynamics Vol. 113; no. 14; pp. 18771 - 18793
• Main Authors: Ye, Chunhui, Tong, Xiaojun, Zhang, Miao, Wang, Zhu
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.07.2025
• Subjects: Applications of Nonlinear Dynamics and Chaos Theory; Classical Mechanics; Control; Dynamical Systems; Physics; Physics and Astronomy; Statistical Physics and Dynamical Systems; Vibration; Non-degeneracy chaos; Compressive sensing; Data hiding; Encrypted images
• ISSN: 0924-090X; 1573-269X
• DOI: 10.1007/s11071-025-11079-2

In recent years, the application of information hiding technology in cloud storage has received much attention. However, most existing schemes struggle to strike a balance between embedding capacity and reconstructed image quality. This paper introduces a high-capacity encrypted image information hiding scheme, which is predicated on a four-dimensional non-degenerate chaotic system and hybrid coding. The proposed hybrid coding method compresses the lower bit planes, which contain less visual information, using compressed sensing technology. This process reduces their occupied space, thereby augmenting the embedding capacity. For the higher bit planes, which carry more visual information, lossless compression is employed using high zero-value continuous bit plane coding technology. This approach preserves crucial contour and brightness information of the image, thereby enhancing the quality of the reconstructed image. Additionally, the image and embedded information are encrypted using a pseudo-random sequence generated by the four-dimensional non-degenerate chaotic system, which demonstrates strong randomness. Experimental results indicate that this scheme can provide superior security while maintaining a balance between embedding capacity and reconstructed image quality. In scenarios where high reconstructed image quality is paramount, the average embedding rates on the BOSSbase and BOWS-2 datasets reach 4.31 and 4.24 bpp respectively, thereby outperforming existing schemes. Furthermore, this scheme exhibits resistance to brute force attacks, statistical attacks, and chosen plaintext attacks.