Lexical stress in continuous speech recognition

Rogier C. van Dalen, Pascal Wiggers, Léon J. M. Rothkrantz

Abstract: Human listeners use lexical stress for word segmentation and disambiguation. We look into using lexical stress for large-vocabulary speech recognition for the Dutch language. It appears that beside vowels, consonants should be taken into account. By introducing stressed phonemes, and features for spectral bands and the fundamental frequency, we reduce the word error rate by 2.6%.