基于音频特征的小波域零水印算法

该文提出一种新颖的基于音频内容的零水印算法,利用音频统计特征,选择小波分解信号中的小波近似系数CA3平均值作为参考特征统计量,运用零水印技术构造鲁棒水印,有效解决水印中鲁棒性和不可察觉性两者间的矛盾.实验表明,该算法对常见的信号处理和攻击都具有较强的鲁棒性和不可察觉性.

A DWT Domain Zero - watermark Algorithm Based on Audio's Character

a new zero-watermark algorithm based on audio's statistical character is presented.By using the audio's statistical character,we consult the mean of the approximation coefficients of audio signal wavelet transform and construct the watermark sequence within the zero-watermark scheme.The statistical character is robustness to the signal attacks.By this way,it is efficient to resolve the required trade-off between the robustness to certain watermark attacks and the imperceptible quality.The experimental results demonstrate that the proposed watermarking algorithm is greatly imperceptible and robust against general malicious attacks.