基于分块DCT的自适应水印算法

Adaptive image watermarking algorithm based on block-DCT

近几年来 ,随着因特网的普及和发展 ,数字图像的版权问题越来越引起人们的关注 ,其中数字水印作为保护版权的一种有效手段得到了广泛的注意。文中提出一种自适应水印算法 ,并比较了Gauss和Legendre两种不同的随机阵列的效果。该方法将随机阵列添加到图像子块的DCT次低频系数中去 ,并根据HVS的特点进行自适应处理和视觉掩盖。实验结果表明 ,实现的水印具有不可见性 ,而且对于常见图像压缩和噪声干扰具有很好的健壮性 ,而采用Legendre阵列的算法效果更佳。

Recently, with the popularity and the development of Internet, the copyright of digital image has aroused extensive interests, among which digital watermarking is deemed as an effective way for copyright protecting. In this paper, an adaptive algorithm is presented and two random arrays, Gauss and Legendre array, are adopted and compared. The array is embedded to the sub-low frequency coefficients of the block-DCT, then adaptive processing and visual masking are made based on the characteristics of HVS. Experimental results demonstrate that the watermark is invisible and robust against common image lossy compression and noise, and the Legendre array is better.