声纹鉴定中自动声纹比对阈值设定技术研究与实践

Research and Practice of Automatic Voiceprint Comparison Threshold Setting Technology in Voiceprint Identification

声纹鉴定系统在实际使用过程中依赖鉴定人员个人经验设置比对阈值进行认定,此种静态阈值设定方法不仅带有主观因素而适应性不佳,且无法适应不同音频质量和信道条件造成的影响,降低了声纹鉴定系统的准确性和可靠性。本文针对自动声纹比对阈值设定技术进行了深入研究与实践,通过使用准确率、召回率和错误拒绝率等不同的评估指标来评估不同阈值设定策略的性能,本研究的结果对于声纹鉴定系统的改进和优化具有重要的意义。

In the actual use of the voiceprint identification system,the affirmation is concluded with a comparison threshold set based on on the personal experience of the identification personnel.This static threshold setting method not only has subjective factors and poor adaptability,but also cannot adapt to the influence of different audio quality and channel conditions,which reduces the accuracy and reliability of the voiceprint identification system.In this paper,the automatic voiceprint comparison threshold setting technology is studied and practiced,and the performance of different threshold-setting strategies is evaluated by using different evaluation indicators such as accuracy,recall rate and false rejection rate.The results of this study are of great significance for the improvement and optimization of the voiceprint identification system.