说话人计算机识别系统的设计与实现

介绍了说话人计算机识别系统的基本原理和研究现状，通过对多种可能方案的比较和选择，设计出了一个说话人识别系统，该系统用汉语的6个单韵母作为录入语音，使用了包括短时平均幅度、短时平均过零率、倒频域分析技术和线性预测分析技术等来完成语音的端点检测和语音个性化特征参数的提取，在Windows环境下用MATLAB完成了系统的开发，使系统短小精悍，并取得了较好的实测效果。

Design and implementation of a speaker recognition system

After a various of comparison among possible speech processing approaches, we present and developed a speaker recognition system(SRS). The six Chinese single vowels are selected as the test elements of our models. The computer speech analysis techniques adopted in this system include short\|time power and short\|time average zero\|cross rate test, cepstrum analysis and linear prediction coding, which are used to detect the start\|end points of speech and to get the characteristic parameters of the speaker. This system is implemented under MATLAB and Windows environment and has perfect recognition results.