Abstract

Automatic Speech Recognition (ASR) is an essential component for applications based on human-machine interfaces. Even after years of development, ASR remnants as one of the crucial research challenges like language variability, vocabulary size and noise. There is a need to develop Human-machine interface in native languages. In this regard, review of existing research on speech recognition is supportive for carrying out further work. Intend of Automatic speech recognition system requires cautious interest to the issues such as category of speech types, feature extraction, pattern classification etc. Here the paper presents a study on typology in ASR, the various phases involved, a brief description on each phase, basic techniques that make up automating speech recognition and different approaches to gain ASR. As an account of the brief study the paper shows enhanced precision results and good accuracy. The paper also displays a swot on speech recognition applications evoking research developments.

Paper Details