基于小波域的稳健性数字图像水印优化算法

An optimized algorithm of robust digital image watermarking based on wavelet transform domain

通过对离散小波变换域(DWT)下小波图像系数的分布特点和不同系数区对嵌入水印的敏感性分析,水印应当首先嵌入小波图像低频系数,若有剩余,再按小波图像频带重要性的排序嵌入高频带.同时在满足不可见性的前提下,应合理分配水印信号的能量,尽可能地提高局部嵌入水印分量的强度.据此,提出了一种基于遗传算法的小波域优化水印嵌入算法.实验结果表明,应用本算法所实现的水印不可见,并对常见的图像处理和噪声干扰具有很好的稳健性.

Through analyzing the distributive features of wavelet image coefficients in the domain of Discrete Wavelet Transform （DWT） and the sensitivity of different coefficient areas to embedded water- marks, it is proposed that watermarks should be imbedded in low-frequency subband coefficients of wavelet images, in case of residues, they should be inserted into the high-frequency subband in order of importance. Furthermore, it suggests that the energy of watermark signals should be reasonably as- signed, and the intensity of watermark weight in some areas should be enhanced maximally, on the premise that invisibility is guaranteed. On these grounds, the wavelet domain optimal watermark imbed- cling algorithm is put forward based on the genetic algorithm. Experimental results show that water- marks generated with the proposed algorithm are invisible, and robust in image processing and against noise interference.