基于整数小波变换的数字水印算法

数字水印技术作为数字产品版权保护的一项新技术,已受到越来越多的关注.鉴于小波变换的多分辨率分析与人眼的视觉特性(Human Visual System,HVS)相一致,提出了一种基于整数小波变换(Integer Wavelet Transform,IWT)和人类视觉系统特性的灰度图像数字水印算法.根据原始灰度图像的整数小波系数高、低频分量的特点,选择原始图像的中频分量嵌入水印,利用HVS特性将二值水印图像加密后嵌入到中频分量的整数小波系数中去,更好地兼顾了水印不可见性与鲁棒性之间的矛盾.采用整数小波变换可以有效地避免运算过程中引入量化误差的缺陷,从而达到精度上的要求,满足图像的精确分解与重构.实验证明,对锐化、JPEG2000压缩和旋转等图像处理均具有很强的抵抗能力,复杂度较低,实用性较强.

A Digital Watermarking Algorithm for Image Based on IWT

This paper proposes a practical algorithm of digital watermark based on IWT and HVS. To ensure the security, the watermark was rotated and scrambled before embedding, and we can get a private key using filter banks in the process. Then according to the different characteristics of the high and low frequency components of the wavelet coefficients of the original image, we select the proper position to embed the watermark. Wavelet transforms match the HVS model, thus it can balance the contradiction between transparency and robustness of the digital watermark well. Because of the storing characteristics of the image, wavelet transforms that map integers to integers will not introduce quantifying error during calculation. The experiment results demonstrate its better resistance, lower complexity and higher practicability.