【摘要】 研究的说话人识别系统,采用能够反映人对语音的感知特性的线性预测(LPC)倒谱参数作为特征参数,同时对特征参数各维分量的识别能力进行定量分析,采用一种新的加权方法进行矢量量化,在此方法下系统取得识别率很高的效果,而且计算量和存储量都比较低.更多还原

【Abstract】 LPCCEP coefficients are generally used in the speaker recognition system.A new weighted method based on considering the recognition capability for each heft of feature parameters is applied in the speaker recognition system.The system with the new method on vector quantifization has obtained better recognition rate;and the requirement of computation and storage is quite low.更多还原