摘要
动态时间规整(DynamicTimeWarping)是语音识别中的一种经典算法,由于算法简单又有效,因此在实现孤立词识别系统中获得了广泛的应用,但仍然存在一些不足,本文从提高识别率和识别速度入手提出对DTW算法进行改进研究,以达到更满意的语音识别效果。
Dynamic Time Warping is a kind of classical programming in speech recognition system ,because the programming is simple and valid ,therefore ,it is achieved extensive application in speech recognition of single word,but there are some questions, the paper raises improved ways from high distinguishable rate and speed ,so that it will attain satisfied effect in speech recognition .
出处
《微计算机信息》
北大核心
2006年第02Z期267-269,共3页
Control & Automation
关键词
语音识别
端点检测
DTW算法
Speech Recognition Beginning Inspect DTW Programming
