中文摘要：

针对信息隐藏技术的抗剪切性和不可见性，提出了一种基于载体自修复系统的信息隐藏算法。该算法将载体图像进行GHM多小波变换，选取高能量子图作为隐藏区域，对其进行置乱处理消除相关性；根据位平面理论进行位平面分解，为保证算法的不可见性，选取低四位进行奇异值分解；根据奇异值的稳定性，将秘密信息嵌入奇异值，增强了算法的鲁棒性。在提取秘密信息时，对受损的含密图像进行自修复处理，减小与原含密图像的差异，保证了奇异值的稳定性。实验表明，该算法较其他算法有较强的鲁棒性和较好的不可见性，特别是抗剪切性。

英文摘要：

Aiming at the anti-cutting and invisibility, an information hiding algorithm based on carrier self- healing system is proposed. This algorithm first implements GHM multi-wavelet transform of the carrier image and then chooses high-energy sub-images as the embedding regions. Those high-energy sub- images are scrambled to eliminate the correlation and then decomposed in line with the bit plane rules. To ensure invisibility of the algorithm, the first four bit planes are chosen and processed with singular value decomposition. Finally, the confidential information is embedded in the singular values in accordance with their stability, thus enhancing the robustness of the algorithm. For the extraction, the damaged stego images are self-healed, thus to reduce the difference between those damaged stego images and the original images and ensure the stability of singular values. The experiments indicate that this algorithm is better than other traditional algorithms in robustness, invisibility, and in particular, the anti-cutting ability.