Speech recognition using syllable and pseudo articulatory features modeling

This paper presents an account of the use of syllable structure as the basis for a novel approach to speech recognition. It is demonstrated that working with syllables provides the basis for linguistically motivated speech recognition using the previously reported notion of the pseudo-articulatory representation (PAR). Speech recognition experiments show that a recognition model developed on the basis of a single speaker can be successfully applied to recognition of speech produced by a different speaker. Therefore, the approach reported demonstrates the potential for speaker independent recognition. In general, this research leads to a more plausible style of automatic speech recognition and contributes to knowledge of phonetics and speech behavior.