Improved lexicon modeling for continuous speech recognition
Seong Jin Yun, Yung Hwan Oh, Gyung Chul Shin · 2002
We propose the stochastic lexicon model which represents the pronunciation variations to optimally cope with the continuous speech recognizer. In this lexicon model, the baseform of words are represented by subword states and the probability distribution of subwords as a hidden Markov model. Also, the proposed approach can be applied to a system employing non-linguistic recognition units and the lexicon is automatically trained from training utterances. In speaker independent speech recognition tests using a 3000 word continuous speech database, the proposed system improves the word accuracy by about 27.8% and the sentence accuracy by about 22.4%.