Vocabulary Speech Recognition via Telephone

Abstract

Abstract The Speech recognition one of important issue because the technique of the systems of speech recognition collection text very expensive. This paper we achieved the needed steps to achieve the independent speech vocabulary recognition and we compare our result with actual speech recognition via telephone. We preformed two sets of experiments, 1) comparing Independent Vocabulary (VI) as command recognition, and dependent vocabulary (VD) systems also word systems. 2) used 42 cites name to recognised which is independently recognised (85% W.Err) which is shown much better result than the (VI) done before by other reports. The context dependence of phonetic models is defined by phonological features rather than by triphone contexts. We achieved batter result for the independent vocabulary, while we recognised there is more work need, we recognise areas for development that are expected to drive to the goal. The goal is to give users a flexible speech recognition system that can be easily modified for different applications with a minimum amount of training and modification effort.