基于音频特征和低频系数较小值的水印算法

摘要

This paper presents a digital watermarking algorithm based on audio features and the relatively smaller values of the low frequency of DWT coefficients. The proposed algorithm analyzes the zero-cross ratio and the short-time energy of frame to select the appropriate threshold for initial discarding the high-frequency signal components of the audio frame, selecting the audio frame to be processed. The selected audio frames are stitched together and perform discrete wavelet transform, choosing the low-frequency coefficients and segmenting, using the sum binary value of the neighboring three watermark bit combination to determine the location of the watermark embedding, and make its value smaller. The proposed algorithm for blind watermark detection reduces the extraction time complexity of low-frequency components through the analysis of audio features, and improves the robustness of the watermark.%提出一种基于音频特征和DWT低频系数相对较小值的数字水印算法.分析音频帧的过零率及短时能量,选取适当的阈值初步舍弃表明信号中高频信号成分的音频帧,筛选出待处理的音频帧.将选定的音频帧拼接在一起进行小波变换,选取低频系数并分段,通过相邻3个水印比特组合的二进制之和确定水印在该段中的嵌入位置,将系数修改为相邻系数中较小的值.实验结果表明,该算法通过对音频特征的分析,能降低提取低频分量的时间复杂度,实现水印信息的盲检测,提高水印的鲁棒性.