一种新的数字语音通信中说话人身份认证方法  

A novel method of speaker identification in digital voice communication

作　　者：唐升[1] 侯榆青[1] 李岩[1] 克兢[1] 

机构地区：[1]西北大学信息科学与技术学院,陕西西安710069

出　　处：《西北大学学报（自然科学版）》2006年第1期46-49,共4页Journal of Northwest University（Natural Science Edition）

基　　金：陕西省自然科学基金资助项目(2003F23)

摘　　要：目的为保证网络信息交互的安全,在数字语音通信中对说话人进行身份认证。方法利用回声隐藏技术,在语音信号中嵌入表征说话人身份的数字签名。结果实验证明,引入的回声信号听觉上不能感知。通过倒谱分析,能够完整地从隐写语音信号中提取出数字签名。结论是一种有效的说话人认证方法,具有广泛的应用前景。Aim To identify the ID of the speaker in digital voice communication in order to assure the security of information which is transmited in the network. Methods Embeding a digital signature that stands for the ID of the speaker into the sequence of voice signal on the basic of the principle of echo hiding. Results The result of the experiment showed that the introduction of the echo has not deteriorated the voice signal＇s perceptual quality. The digital signature can be extracted from the embedded voice signal by the method of cepstrum analyzing. Conelusion This is an effective method of speaker identification, and has a broad application field.

关 键 词：说话人认证 回声隐藏 数字签名 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]