On the Implementation of a Speech Recognizer for Romanian Language Based on Hidden Markov Models

Over the last decades there has been made a considerable progress in the field of automatic speech recognition. New methods and technologies have been developed and we can say that we have currently reached a state-of-the-art in this field. Most of the latest speech recognizers are based on hidden Markov models since it has been proved that this is the best method to model speech. This paper presents the steps involved in building this kind of a speech recognition system, together with a few results obtained in the recognition of isolated words (digits from zero to nine) in Romanian language