대한음성학회:학술대회논문집 (Proceedings of the KSPS conference)
- 대한음성학회 2007년도 한국음성과학회 공동학술대회 발표논문집
- /
- Pages.131-134
- /
- 2007
다단계 인식기반의 POI 인식기 개발
Multi-stage Recognition for POI
- 전형배 (한국전자통신연구원 음성/언어정보연구센터) ;
- 황규웅 (한국전자통신연구원 음성/언어정보연구센터) ;
- 정훈 (한국전자통신연구원 음성/언어정보연구센터) ;
- 김승희 (한국전자통신연구원 음성/언어정보연구센터) ;
- 박준 (한국전자통신연구원 음성/언어정보연구센터) ;
- 이윤근 (한국전자통신연구원 음성/언어정보연구센터)
- Jeon, Hyung-Bae (Speech/Language Information Research Center, ETRI) ;
- Hwang, Kyu-Woong (Speech/Language Information Research Center, ETRI) ;
- Chung, Hoon (Speech/Language Information Research Center, ETRI) ;
- Kim, Seung-Hi (Speech/Language Information Research Center, ETRI) ;
- Park, Jun (Speech/Language Information Research Center, ETRI) ;
- Lee, Yun-Keun (Speech/Language Information Research Center, ETRI)
- 발행 : 2007.05.18
초록
We propose a multi-stage recognizer architecture that reduces the computation load and makes fast recognizer. To improve performance of baseline multi-stage recognizer, we introduced new feature. We used confidence vector for each phone segment instead of best phoneme sequence. The multi-stage recognizer with new feature has better performance on n-best and has more robustness.
키워드