摘要
本文介绍亍脉动阵列算法的原理,给出了语音识别中进行模式匹配使用的动态规划算法和概率匹配算法的实现。这种结构对实时的大词汇量的孤立词或连续语音识别是十分有效的。
The principles of the systolic arrays and its realization using DTW algorithm and probability matching lgorithm for pattern matching in human speech recognition are presented. It is efficient for solving the real-time isolated or connected words recognition problems with great number of vocabularies.
出处
《南京邮电学院学报》
北大核心
1990年第2期7-24,共18页
Journal of Nanjing University of Posts and Telecommunications(Natural Science)
关键词
语言识别
脉动阵列
算法
Speech recognition
Systolic Arrays
Algorithm
