Hervé Bourlard
In this report, we present preliminary experiments towards automatic inference and evaluation of pronunciation models based on multiple utterances of each lexicon word and their given baseline pronunciation model (baseform phonetic transcription). In the p ...
IDIAP2001