人体生理信息的语音表征在声纹鉴定中的应用

On the Application of Voice Representation of Human Physiological Information in the Identification of Voiceprint

语音信号可以传递多种信息。研究发现,生理信息(如体型、年龄及病理等)在语音信号中具有一定的表征。基于生理信息和语音特征的相关分析,现有技术已能够在一定程度上实现基于语音信号的说话人生理信息(如舌体模型和帕金森病等)重现。在声纹鉴定中,若能将生理信息的语音表征纳为参考,不仅能进一步了解说话人,丰富鉴定的证据来源,还能为案件侦破提供一定的方向。因此,人体生理信息的语音表征对声纹鉴定具有潜在的应用价值。

Speech signals can convey a variety of information.Previous work has found that physiological information(e.g.,body size,age,pathology,etc.)is represented in the speech signal to some extent.Based on the correlation analysis of physiological information and speech feature,existing techniques have been able to reproduce the speaker’s physiological information(e.g.,tongue body model,Parkinson’s disease,etc.)based on speech signals to some extent.In voice identification,if the speech representation of physiological information can be incorporated as a reference,it can not only further improve the understanding of the speaker and enrich the source of evidence for identification,but also provide directions for case detection.Therefore,the acoustic representation of human physiological information has potential application value for voice identification.