一种引入帧间相关信息的HMM语音识别方法

赵力; 邹采荣; 吴镇扬

一种引入帧间相关信息的HMM语音识别方法

计量
- 文章访问数: 2064
- HTML全文浏览量: 126
- PDF下载量: 448
- 被引次数: 0
出版历程
- 收稿日期: 1999-06-11
- 修回日期: 2000-04-23
- 刊出日期: 2001-04-19

A METHOD OF HMM SPEECH RECOGNITION INTRODUCED INTER-FRAME CORRELATION

摘要

摘要: 该文提出了一种基于复数帧段输入HMM的语音识别方法,它采用相继的复数帧组成的特征参数向量作为语音识别HMM的输入,能有效地在语音识别HMM中引入帧间相关信息。为了进一步改善复数帧段输入HMM的输出概率分布函数,作者还提出了用MGDF和RBF函数作为复数帧段输入HMM的输出概率分布函数的方法。通过对非特定人汉语孤立数字和连续数字语音识别试验,证实了该文提出的引入帧间相关信息方法的有效性。
- 语音识别; 隐马尔可夫模型; 帧间相关信息; 复数帧段输入
Abstract: This paper applies segmental unit into HMM for speech recognition. In this model, several successive frames are combined and treated as an input vector. It expects that segmental unit input HMM would be effective to describe the inter-frame correlation information and has also proposed the MGDF and RBF to further improve output probability function. By comparing them with the traditional HMMs based on their speech recognition performance rates through the experiments of speaker-independent spoken digit (isolated/connected) recognition,the validity of the proposed appraoch could be verified.

HTML全文

参考文献(1)

V.N. Gupta, M. Lennig, P. Mermelstein, Integration of acoustic information in a large vocabulary word recognizer, ICASSP-87, Dallas, USA, 1987.2, 697-700.[2]坪香英一,ニュ-ラルネツト驱动型HMM, IEICE, Technical Report, 1989, SP89-83, 33-41.[3]L. Deng, M. Aksmanoric, X. Sun, C. F. J. Wu, Speech recognition using hidden Markov models with polynomial regression functions as stationary states, IEEE Trans.[J]. on Speech Audio Processing.1994,(4):507-[4]C.J. Wellekens, Explicit correlation in hidden Maarkov model with optimized inter-frame dependence, ICASSP-95, Detroit, USA, 1995.1,209-212.[5]相川清明,河原英纪,顺向マスキングの时间周波数特性を模拟した动的ケブストラムを用いた音韵识别,日本通信电子学会论文志(A),1993,J76-A(11),1514-1521.[6]井手和之,牧野正三,时间-周波数を用いた无声破裂音の识别,日本音响学会论文志,1982,39(5),321-329.[7]M. Ostendorf, S. Roukos, A stochastic segment model for phoneme-based continuous speech recognition, IEEE Trans. on Acoust..[J]. Speech Signal Processing.1989,ASSP-37(12):1857-[8]T. Wakabayashi, S. Tsuruokaet, ed al., On the size and variable transformation of feature vector for handwritten character, IEICE, J76-D- Ⅱ (12), 2495-2503.[9]L. Zhao, H. Suzuki, S. Nakagawa, A comparison study of probability functions in HMMs through spoken digit recognition, IEICE, TRANS.INF and SYST., 1995, E78-D(6), 669-675.[10]S. Nakagawa, Estimation of probability density function and a posteriori probability and evaluation by vowel recognition, IEICE, Technical Report, 1992, SP92-24, 61-72.

施引文献

资源附件(0)

访问统计