Abstract: In view of the contradiction between capability, invisibility and robustness of the existing information hiding algorithms, a preprocessing algorithm for digital image based on CL (Chui-Lian) multi-wavelet transform and Combination Bit Plane (CBP) was proposed. Then the digital image preprocessed by CL multi-wavelet and CBP was taken as the cover image to embed secret information for confidential communication and image sharing. CL multi-wavelet transform could divide the cover image into four lowest resolution sub-images with different energy level. And CBP method could analyze the above four sub-images into different bit planes as the final embedding regions. During the hiding procedure, robust information, secret information and fragile information could be embedded according to the energy and robustness characteristics of the embedding regions. The experimental results show that the robustness against several common attacks described in this paper has certain enhancement compared with other two methods when the embedding rate is 25%. The Peak Signal-to-Noise Ratio (PSNR) is increased by 37.16% and 20.00% respectively compared with Discrete Cosine Transform-Least Significant Bit (DCT-LSB) and Discrete Wavelet Transform -Least Significant Bit (DWT-LSB) algorithm.

Key words: information hiding, confidential communication, carrier preprocessing, CL (Chui-Lian) multi-wavelet transform, Combination Bit Plane (CBP) theory

摘要： 针对信息隐藏算法容量和不可见性与鲁棒性矛盾的特点，提出了一种新的基于Chui-Lian(CL)多小波变换和组合位平面(CBP)理论的数字图像载体预处理算法，并将该算法用于以数字图像为载体的信息隐藏，以实现秘密通信和图像共享。其中，CL多小波变换能够将载体图像划分为4个能量等级不同的最低分辨率子图像，组合位平面理论可将子图像解析为不同的位平面层次，在隐藏秘密信息时，可根据最低分辨率子图和组合位平面的能量等级由高到低分别嵌入鲁棒信息、秘密信息和脆弱信息。实验结果表明，在25%的嵌入率时，与基于离散余弦变换和最低有效位方法的信息隐藏算法（DCT-LSB）、基于离散小波变换和最低有效位方法的信息隐藏算法（DWT-LSB）相比，所提算法针对若干常见攻击的鲁棒性有一定程度提升，峰值信噪比（PSNR）分别提高了37.16%和20.00%。

关键词: 信息隐藏, 秘密通信, 载体预处理, CL多小波变换, 组合位平面理论