Forschungsgebiete
- Visualisation NavCARS-EVAR
- Robotics Laboratory
- High-Accuracy Head Tracking
- Bewegungskompensation
- Ultraschallgeführte Strahlenchirurgie
- Modellierung von Herzbewegungen
- Neurologische Modellierungen
- Autonome Elektrofahrzeuge
- Ziel-basierendes lebenslanges
autonomes Lernen - Transkranielle Elektrostimulation
Ehemaligen Forschungsgebiete
Recognition of Non-Native Speech Using Dynamic Phoneme Lattice Processing
| Type of publication: | Artikel in einem Konferenzbericht |
| Buchtitel: | Acoustic Society of Japan, spring meeting 2002 |
| Jahr: | 2002 |
| Ort: | Yokohama, Japan |
| Abriss: | Robustness of speech reconizers to non-native speech input is a difficult problem with rising importance. This paper introduces an approach which deals with the dynamic application of pronunciation rules to a phoneme-lattice. These rules are generated in a separate step. The task is the robust recognition of English spoken by Japanese for hotel reservation, using the Wall-Street-Journal (WSJ) for baseline acoustic modeling. |
| Schlagworte: | |
| Autoren: | |
|
Anhänge
|
|
| [Bib|RIS] | |
|
|
|