期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2012
卷号：3
期号：3
页码：3989-3992
出版社：TechScience Publications
摘要：-Pattern matching is an emerging subject for doing identification of various objects including speech and its parts. Initially these pattern matching techniques has been limited to identifying the speech patterns of the words that are separated with enough silence. It is easy to segment each word and extract its features which machine algorithm like Neural Network can learn and simulate, but if a sufficient interval of silence is not there, then there are connected words between each pause based on persons rate of speech, if for example person has spoken 5 types of words the connected combination might reached to 120 which would ultimately lead to a great challenge for pattern matching algorithm. In this paper we are reviewing such methodologies.
关键词：Dynamic Time Warping (DTW); Hidden Markov Model;(HMM) and Artificial Neural Network (ANN)RMSE
