Survey of Technical Progress in Speech Recognition by Machine over Few Years of Research

Abstract

Abstract: This paper reviews some of various research carried out over the last decade in the area of Automatic Speech Recognition (ASR) and discusses the major themes and advance made in the last decade of research, in order to show the outlook of technology and an appreciation of the fundamental progress that has been achieved in this weighty area of speech communication. Over period of research and development, the accuracy of automatic speech recognition remains one of the important research challenges such as variation of the context, environmental condition, speaker’s variation and poor-quality audio. The design of speech recognition requires careful attention to the following issue: Definition of various types of speech classes, speech representation, techniques, database and performance evaluation. The history, challenges of speech recognition system and various techniques to solve these challenges constructed by various research works have been presented in a chronological order. The objective of this paper is to compare and summarize well know approaches used in various steps of speech recognition system. Keywords: Automatic Speech Recognition, Database, Feature extraction, Performance evaluation, Techniques