Phonetic Distance Measures for Speech Recognition Vocabulary and Grammar Optimization

Abstract

This paper reports on the correlation between word confusion matrices from Word-Error-Rate (WER) experiments and different phonetic distance measures. The investigated phonetic distance measures are based on the minimum-edit-distances between phonetic transcriptions and the distances between Hidden-Markov-Models (HMM). We show that phonetic distance measures are correlated with word confusion. The correlations between word confusion of a speech recognizer and phonetic distance are useful for a speech recognition grammar developer or a spoken dialog system designer in developing efficient grammars and dialogs. Furthermore the measures can be used for evaluating the quality of grammars in terms of phonetic confusability of words/utterances or interpretations. An extension of these measures to grammar optimization is discussed.

Extracted Key Phrases

6 Figures and Tables

Cite this paper

@inproceedings{Pucher2007PhoneticDM, title={Phonetic Distance Measures for Speech Recognition Vocabulary and Grammar Optimization}, author={Michael Pucher and Andreas T{\"{u}rk and Jitendra Ajmera and Natalie Fecher}, year={2007} }