Robust Watermarking Scheme for Encrypted Images Based on Scrambling and Kronecker Compressed Sensing

Data hiding in encrypted images (DHEI) embeds data in the cipher state, which not only protects the confidentiality of the cover image content, but also hides the existence of the secret information, and has a broad application prospect. However, many DHEI schemes pursue high capacity at the expense of robustness. In this letter, we propose a robust watermarking scheme for encrypted images based on scrambling and Kronecker compressed sensing (CS). The scheme makes full use of the encryption and compression characteristics of CS to realize the image encryption and preprocessing at the same time, and utilizes the democratic characteristic of CS, Most Significant Bits (MSBs) replacement, image global scrambling and adjusted uniform quantization to improve the robustness. In addition, the two-dimensional Kronecker CS reduces the computational complexity of the encoding end and increases the encoding efficiency. Experimental results show that our scheme achieves good robustness against various noise attacks and clipping attacks, while maintaining high-capacity embedding and stable recovery.