期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2016
卷号：7
期号：3
页码：1116-1120
出版社：TechScience Publications
摘要：An automatic speech recognition (ASR) system isvery popular and efficient process to converts sound signalsinto text. It is widely used in various applications in marketsuch as speech-to-text, entries of data etc. Accuracy of ASR isone of the big challenge for the researchers because manyvoice inputs have noise and declined channel, it resultsmismatched condition from the point of view of competencyand reliability. In sound recognition, it has three mostimportant approaches as Artificial intelligence, second isPattern recognition and the last one is Acoustic phonetics. Thegoal of this paper is to clear an idea about basics of voicerecognition by differentiating methods and summarizing atheory of what is ASR? And how does it works? in each andevery stages of voice recognition system. It discusses theprogress and development of voice recognition in last fivedecades that has been performed and describes a literaturereview of a system. This paper presents innovative anddistinctive features of a system.
关键词：Automatic speech recognition; Pattern recognition;Language model; Hidden markov model; Linear discriminate;analysis; Mel frequency cepstral coefficient
