一种新的基于变换域的数字音频水印算法

A New Transform Domain-based Digital Audio Watermarking Algorithm

使用Arnold变换对二值黑白水印图像进行预处理,对选定的音频进行五层小波分解,对小波分解后的低频系数进行离散余弦变换(DCT),选取DCT系数中幅值较大的点嵌入水印信息,提出一种基于离散小波变换和离散余弦变换相结合的数字音频水印方案。实验表明:该算法具有较好的透明性,对噪声、重采样、重量化等攻击具有较好的鲁棒性。

In this paper, we used Arnold transform to do pre-processing of binary black and white watermark image, made five-level wavelet decomposition of selected audio and the discrete cosine transform （DCT） of low-frequency coefficient after wavelet decomposition. Also we chose the larger point embedded watermark information of DCT coefficient, and proposed a design of digital audio watermarking combining wavelet transform and DCT.The experiments showed that the algorithm had good robustness under attacks of noise, resample and weight.