Speech Recognition using Artificial Neural Networks

Abstract

As most of the researches on speech recognition (SR) are based on hidden Markov models (HMM), the main theme of this paper is the recognition of Arabic sounds using artificial neural networks. Despite the fact that Arabic is a language that is spoken by millions of people, and it is the sixth (K. Kirchhoff and J. Bilmes, 2002) spoken language in the world, we have faced a scarcity of researches in Arabic language recognition during the preparation of this paper. Speech recognition systems will become more used as they started to replace some of the functions normally accomplished with a keyboard, these and many other reasons encouraged us to continue in this field