摘要
为了将数字音频水印技术与音频内容相结合,基于人声和乐器的基频稳定的特点,从音乐中提取出基频并据此设计一种鲁棒零水印方法。首先通过正规化分谐波叠加算法提取基频,然后采用K-means算法对基频特征进行编码加强其稳定性,最后与水印图像异或生成零水印序列。此外,发生恶意窜改时,对比零水印中的基频信息得到不一致的部分,即可确定恶意窜改区域。实验结果表明,此算法在常规攻击及抖动攻下均具有良好的稳定性,且可实现窜改检测。
In order to combine the digital audio watermark technology with the audio content,because of the stability of the fundamental frequency of the human voice and musical instruments,this paper designed a robust audio zero-watermarking method based on the fundamental frequency of music. Firstly,the normalized subharmonic summation algorithm extracted the fundamental frequency from each frame of the audio signal. Then,the K-means algorithm encoded the extracted fundamental frequencies so as to further improve the stability of the feature. Finally,the encoded fundamental frequencies XOR with the watermark image to generate a zero watermark sequence. Besides,when the audio is maliciously tampered,it can get the inconsistent part of the audio signal by comparing the fundamental frequency information contained in the zero watermark sequence. At the inconsistent section,the algorithm can detect the maliciously tampered area. The experiment results show that the proposed algorithm has good robustness against common attacks and jitter attacks,and the results also prove that the algorithm can detect and locate the tampering.
作者
郑梦怡
李晨
田丽华
Zheng Mengyi;Li Chen;Tian Lihua(School of Software Engineering,Xi’an Jiaotong University,Xi’an 710049,China)
出处
《计算机应用研究》
CSCD
北大核心
2019年第2期538-542,共5页
Application Research of Computers
基金
国家自然科学基金资助项目(61403302)
中央高校基本科研业务费专项资金资助项目(XJJ2016029)
关键词
数字音频水印
鲁棒零水印
正规化分谐波叠加
基频
窜改检测
digital audio watermarking
robust zero watermark
normalized subharmonic summation
fundamental frequency
tamper detection
