摘要
文章介绍了一种非统计地分析具有高斯分布样本的模式之间混淆度的方法,并将此种方法应用在语音识别中。实验表明,由此种方法得到的模式之间的混淆度信息对提高模式识别系统的性能有很大的帮助。
In this paper,a non-statistic method of analyzing the confusion between patterns with Gauss samples is in-troduced.And the experiment of it is carried out in speech recognition.It shows in the result that the confusion between two patterns calculated by this method is of great help to recognition work.
出处
《计算机工程与应用》
CSCD
北大核心
2002年第18期11-13,共3页
Computer Engineering and Applications
基金
清华大学"985"重大项目(编号:985校-22-攻关-06)
