State-based bilingual model modification for nonnative speech recognition

Yonghong YanDigital 2009

Abstract

The performance of automatic speech recognition decreases drastically for nonnative speakers,especially those who are just beginning to learn foreign language or who have heavy accents.A novel bilingual model modification approach is presented to improve nonnative speech recognition accuracy.Each state of baseline nonnative acoustic model is modified with several candidate states from the auxiliary acoustic model,which is trained by speakers' mother tongue.State mapping criterion and n-best candidates are investigated.Using this bilingual model modification approach,compared to the nonnative acoustic model which has already been well trained by adaptation technique MAP,the phrase error rate further is reduced by 11.7%relatively,while only a small relative increase on real time factor occurs.
BibTeX
@inproceedings{yan2009state,
  author = {Yonghong Yan},
  title = {{State-based bilingual model modification for nonnative speech recognition}},
  booktitle = {Digital},
  year = {2009},
}

Loading related content...

Figure