语音伪装方法及其防御对策综述  

Voice disguise and its countermeasures for anti-disguise

作　　者：郑琳琳 孙蒙 张雄伟 潘志欣 Zheng Linlin;Sun Meng;Zhang Xiongwei;Pan Zhixin(Army Engineering University of PLA,Nanjing 210007,China;Naval University of Engineering,Wuhan 430000,China)

机构地区：[1]陆军工程大学,江苏南京210007 [2]海军工程大学,湖北武汉430000

出　　处：《信息技术与网络安全》2020年第8期33-42,共10页Information Technology and Network Security

基　　金：国家自然科学基金(61471394);江苏省优秀青年基金(BK20180080)。

摘　　要：语音伪装是指以隐藏说话人身份为目的对说话人的个性特征进行的改变。近年来,随着智能语音交互技术和声纹认证产品的快速发展,语音伪装被应用到语音产品隐私保护中,但是也被不法分子用以实施违法犯罪行为。因此,语音伪装成为目前语音处理和信息安全领域非常有实用意义的研究问题。在简要介绍语音伪装的典型模型和方法的基础上,梳理了语音伪装威胁量化评估方案,总结了近几年伪装语音的防御对策,分析了目前防御对策中存在的问题及技术难点,并对语音伪装的未来研究发展方向进行了展望。Voice disguise refers to the change of the personality characteristics of the speaker for the purpose of hiding the speaker′s identity.In recent years,with the rapid development of intelligent voice interaction technology and voiceprint authentication products,voice disguise has been applied to the privacy protection of voice products,but it has also been used by criminals to commit illegal crimes.Therefore,voice disguise has become a very practical research issue in the field of voice processing and information security.On the basis of introducing the typical models and methods of voice disguise,the methods of quantifying voice disguise threat assessment methods are summarized in this paper,and the defense countermeasures of disguise voice in recent years are introduced.Then,the problems and technical difficulties still existing in defense countermeasures are analyzed.Finally,the prospects for the research development direction of voice disguise are given.

关 键 词：语音伪装 声纹识别 自动说话人确认 语音变换 

分 类 号：TN912[电子电信—通信与信息系统]