Abstract
As an effort to make prosody useful in spontaneous speech recognition, we adopt a quasi-continuous prosodic annotation and accordingly design a prosody-dependent acoustic model to improve ASR performances. We propose a variable-parameter Hidden Markov Models, modeling the mean vector as a function of the prosody variable through a polynomial regression model. The prosodically-adapted acoustic models are used to re-score the N-best output from a standard ASR, according to the prosody variable assigned by an automatic prosody detector. Experiments on the Buckeye corpus demonstrate the effectiveness of our approach.
| Original language | English (US) |
|---|---|
| Title of host publication | 5th International Conference on Speech Prosody 2010 |
| Publisher | International Speech Communications Association |
| ISBN (Electronic) | 9780000000002 |
| State | Published - 2010 |
| Externally published | Yes |
| Event | 5th International Conference on Speech Prosody: Every Language, Every Style, SP 2010 - Chicago, United States Duration: May 10 2010 → May 14 2010 |
Publication series
| Name | Proceedings of the International Conference on Speech Prosody |
|---|---|
| ISSN (Print) | 2333-2042 |
Conference
| Conference | 5th International Conference on Speech Prosody: Every Language, Every Style, SP 2010 |
|---|---|
| Country/Territory | United States |
| City | Chicago |
| Period | 5/10/10 → 5/14/10 |
Funding
We thank J.H. Chu and X. Zhou for helpful discussions. This work was funded by the National Science Foundation Grant CCF 04-26627.
Keywords
- Prosody-dependent ASR
- Re-scoring
- Variable parameter HMM
ASJC Scopus subject areas
- Language and Linguistics
- Linguistics and Language