Proceedings of the KSPS conference (대한음성학회:학술대회논문집)

The Korean Society Of Phonetic Sciences And Speech Technology (대한음성학회)
권호 표지

A VQ Codebook Design Based on Phonetic Distribution for Distributed Speech Recognition

분산 음성인식 시스템의 성능향상을 위한 음소 빈도 비율에 기반한 VQ 코드북 설계

  • Oh Yoo-Rhee (Dept. of Information and Communications, Gwangju Institute of Science and Technology) ;
  • Yoon Jae-Sam (Dept. of Information and Communications, Gwangju Institute of Science and Technology) ;
  • Lee Gil-Ho (CS Management Center, Samsung Electronics) ;
  • Kim Hong-Kook (Dept. of Information and Communications, Gwangju Institute of Science and Technology) ;
  • Ryu Chang-Sun (Speech Research Division, Spoken Language Research Department, Advanced Technology Laboratory Korea Telecom) ;
  • Koo Myoung-Wa (Speech Research Division, Spoken Language Research Department, Advanced Technology Laboratory Korea Telecom)
  • 오유리 (광주과학기술원 정보통신공학과) ;
  • 윤재삼 (광주과학기술원 정보통신공학과) ;
  • 이길호 (삼성전자 CS 경영센터) ;
  • 김홍국 (광주과학기술원 정보통신공학과) ;
  • 류창선 (KT 음성언어연구부) ;
  • 구명완 (KT 음성언어연구부)
  • Published : 2006.05.01
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we propose a VQ codebook design of speech recognition feature parameters in order to improve the performance of a distributed speech recognition system. For the context-dependent HMMs, a VQ codebook should be correlated with phonetic distributions in the training data for HMMs. Thus, we focus on a selection method of training data based on phonetic distribution instead of using all the training data for an efficient VQ codebook design. From the speech recognition experiments using the Aurora 4 database, the distributed speech recognition system employing a VQ codebook designed by the proposed method reduced the word error rate (WER) by 10% when compared with that using a VQ codebook trained with the whole training data.

Keywords