基于参数寻优的立体声鲁棒水印算法  

Stereo robust watermark algorithm based on parameter optimization

作　　者：薛一鸣[1] 张金雨 陈波涛 杨正洪[3] XUE Yiming;ZHANG Jinyu;CHEN Botao;YANG Zhenghong(College of Information and Electrical Engineering,China Agricultural University,Beijing 100083,China;Beijing China Electronics Huada Electronic Design Company,Beijing 102209,China;College of Science,China Agricultural University,Beijing 100083,China)

机构地区：[1]中国农业大学信息与电气工程学院,北京100083 [2]北京中电华大电子设计有限责任公司,北京102209 [3]中国农业大学理学院,北京100083

出　　处：《通信学报》2023年第7期149-158,共10页Journal on Communications

基　　金：国家自然科学基金资助项目(No.62272463);国家重点研发计划基金资助项目(No.2021QY2312)。

摘　　要：针对网络音频数据非法传播、隐私泄露等问题,提出了一种基于参数寻优的立体声鲁棒水印算法。首先利用立体声2个声道之间的相对关系对攻击的不敏感性,以2个声道的相对失真最小为目标,优化嵌入过程,修改奇数段的奇异值比值以嵌入水印,提高音频文件的感知质量。其次,针对不同音频文件,提出一种参数寻优算法用于寻找对每个音频文件最优的参数,更好地平衡了算法的不可感知性和鲁棒性。仿真实验结果证明,所提算法具有较好的不可感知性,并且对常规信号处理均具有较好的鲁棒性。Aiming at issues such as illegal dissemination of network audio data and privacy leakage,a stereo robust watermark algorithm based on parameter optimization was proposed.Firstly,the watermark was embedded by modifying the singular value ratio of odd segment based on the insensitivity of stereo channels to attack,and aiming at minimizing the relative distortion of the two channels.The embedding process was optimized and the perceived quality of the audio was improved.Then,a parameter optimization method was proposed to find optimal parameters on each audio file,and the imperceptibility and robustness were well balanced.The simulation results show that the proposed algorithm has good imperceptibility and robustness against conventional signal processing.

关 键 词：立体声音频 音频水印 参数寻优 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]