Arguably the most important part of automatically assessing a new readers literacy is in verifying his pronunciation of read-aloud target words. But the pronunciation evaluation task is especially difficult in children, non-native speakers, and pre-literates. Traditional likelihood ratio thresholding methods do not generalize easily, and even expert human evaluators do not always agree on what constitutes an acceptable pronunciation. We propose new recognition- and alignment-based features in a decision tree classification framework, along with the use of prior linguistic information and human perceptual evaluations. Our classification methods demonstrate a 91% agreement with the voted results of 20 human evaluators who agree among themselves 85% of the time.
Cite as: Tepperman, J., Silva, J., Kazemzadeh, A., You, H., Lee, S., Alwan, A., Narayanan, S. (2006) Pronunciation verification of children²s speech for automatic literacy assessment. Proc. Interspeech 2006, paper 1814-Tue1WeS.8, doi: 10.21437/Interspeech.2006-286
@inproceedings{tepperman06_interspeech, author={Joseph Tepperman and Jorge Silva and Abe Kazemzadeh and Hong You and Sungbok Lee and Abeer Alwan and Shrikanth Narayanan}, title={{Pronunciation verification of children²s speech for automatic literacy assessment}}, year=2006, booktitle={Proc. Interspeech 2006}, pages={paper 1814-Tue1WeS.8}, doi={10.21437/Interspeech.2006-286} }