节点文献
与文本无关的说话人识别系统的设计
Design and Investigation About a Text-independent Speaker Recognition System
【摘要】 说话人识别是语音信号处理中的重要组成部分,是当前的研究热点之一。详细介绍了说话人识别系统的基本原理和研究现状,并在此基础上对多种可能的特征参数、训练模型和分类方法进行比较和选择,设计出一个识别率高、计算量小的说话人识别系统,识别系统采用MATLAB完成开发,仿真结果显示系统取得了较好的实验结果。
【Abstract】 The speaker recognition is an important part of speech signal processing. It is one of the current research hotspots. The basic principles and researching actuality are introduced in detail. Compared with the various feature parameters, training models and classification methods, a system in a higher recognition rate and lower calculation is designed. The speaker recognition system is developed with MATLAB language. The simulating results indicate the results of the recognition is good.
【关键词】 说话人识别;
特征提取;
模型训练;
分类;
【Key words】 speaker recognition; feature extraction; model training; classification;
【Key words】 speaker recognition; feature extraction; model training; classification;
- 【文献出处】 电声技术 ,Audio Engineering , 编辑部邮箱 ,2006年12期
- 【分类号】TN912.34
- 【被引频次】6
- 【下载频次】156