基于内积能量的小波域自适应水印算法

摘要

针对目前很多小波域水印算法都没有考虑噪声对选择水印嵌入位置的干扰这一问题,提出一种基于内积能量的小波域自适应水印算法.该算法充分考虑人眼视觉系统掩蔽特性,利用梯度内积能量衡量的纹理特征和亮度对噪声的掩饰特性确定水印嵌入强度,将水印分量嵌入到低频系数中,解决了噪声对嵌入位置选择的影响.实验结果表明,该算法在保证水印透明性的同时,对于非几何攻击和几何攻击具有很强的鲁棒性.%Aiming at the issue that many wavelet domain watermarking algorithms at present do not consider the noise interference on the selection of embedding position, a wavelet domain adaptive watermarking algorithm based on inner product energy is put forward. The algorithm fully considers the masking properties of human visual system, uses the texture characteristic measured in gradient inner product energy and the cover characteristic of brightness for noise to determine the watermark embedding strength, embeds the watermark components to low frequency coefficient, thus solves the problem of the noise influence on embedded position selection. Experiment results show that this algorithm guarantees the watermark transparency while has strong robustness against the geometric attacks and non-geometric attacks.