Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
最適化された部分的確率混合共通化を用いる音声認識のための方法および装置
Document Type and Number:
Japanese Patent JP4141495
Kind Code:
B2
Abstract:
In accordance with the invention, a speech recognizer is provided which uses a computationally-feasible method for constructing a set of Hidden Markov Models (HMMs) for speech recognition that utilize a partial and optimal degree of mixture tying. With partially-tied HMMs, improved recognition accuracy of a large vocabulary word corpus as compared to systems that use fully-tied HMMs is achieved with less computational overhead than with a fully untied system. The computationally-feasible technique comprises the steps of determining a cluster of HMM states that share Gaussian components which are close together, developing a subset codebook for those clusters, and recalculating the Gaussians in the codebook to best estimate the clustered states.

Inventors:
Digarakis, Vasilios
Movate, Hay
Application Number:
JP50515196A
Publication Date:
August 27, 2008
Filing Date:
July 13, 1995
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
SRI International
International Classes:
G10L15/06; G10L15/14
Other References:
松岡達雄,「不特定話者音声認識」,電子情報通信学会技術研究報告,1994年5月19日,Vol.94,No.42,SP94-4,p.25-32
Mei-Yuh Hwang et al.,”Subphonetic Modeling with Markov States-Senone”,Proc.of 1992 IEEE ICASSP,1992年3月23日,Vol.1,pI-33-I36
Mei-Yuh Hwang et al.,”Predicting Unseen Triphones with Senones”,Proc.of 1993 IEEE ICASSP,1993年4月27日,Vol.2,pII-311-II314
小坂哲夫 他,「混合連続分布HMM音素モデルの構造自動決定法の検討」,日本音響学会平成4年度秋季研究発表会講演論文集-I-,1992年10月5日,2-1-1,p.79-80
松岡達雄 他,「混合ガウス分布不特定話者HMMをベースとした重み係数による話者適応化法」,日本音響学会平成4年度春季研究発表会講演論文集-I-,1992年3月17日,1-1-6,p.11-12
小坂哲夫 他,「話者混合SSSによる不特定話者音声認識と話者適応」,電子情報通信学会技術研究報告,1992年9月10日,Vol.92,No.207,SP92-52,p.17-24
Attorney, Agent or Firm:
Hidesaku Yamamoto
Takaaki Yasumura
Natsuki Morishita