Applying a Multi-Voice Speech Recognizer to the BMIST Task
Abstract
Speech recognition topics are explored. An orally driven user interface to form filling was developed. Along the way, unsupervised adaptation methods, noise injection and novel enrollment/ model search methods were studied. On the strength of the discovered techniques, an application moved from a research topic to a program of record .
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 2008
- Accession Number
- ADA489824
Entities
People
- Gregory J. Gadbois