Neighboring digits pattern training method in quickly-spoken connected mandarin digits speech recognition

被引：1
作者：
Guo C. [1 ,3 ]
Li R. [2 ]
Fan M. [1 ,3 ]
Liu K. [4 ]
机构：
[1] School of Information Engineering, Zhengzhou University, Zhengzhou, Henan
[2] Henan Provincial Key Lab. on Information Network, Zhengzhou University
[3] School of Information Engineering, Zhengzhou University, Zhengzhou, Henan
[4] Institute of Naval Equipment, Beijing
来源：
Journal of Multimedia | 2011年 / 6卷 / 03期
关键词：
Deletion errors; Discriminability; Hidden Markov model (HMM); Mandarin digit string speech recognition (MDSSR); Neighboring digits pattern; Speaking rate;
D O I：
10.4304/jmm.6.3.300-307
中图分类号：
学科分类号：
摘要：
Deletion errors are most usually occurred in connected Mandarin digit string speech recognition when speaking rate is fast, and are the main reasons leading to the increasing of the recognition error rate and the decline of the recognition accuracy. In this paper, a new training method named neighboring digits pattern is given based on sufficient statistics of recognition errors of the traditional system in order to eliminate most of deletion errors which seriously affect the system recognition rate. The training process is presented and the performance evaluation is given. The result analysis demonstrates that the new method can reduce the deletion errors effectively and improve the system recognition rate from 96.4% to 98.3%. © 2011 ACADEMY PUBLISHER.
引用
收藏
页码：300 / 307
页数：7
相关论文
共 17 条
  • [1] HTK Book, (2006)
  • [2] Shi Y., Liu J., Liu R., Discriminative HMM stream model for mandarin digit string speech recognition, ICSP' 02 Proceedings, pp. 528-531, (2002)
  • [3] Yuanyuan S., Jia L., Runsheng L., Mandarin Digit String Speech Recognition Using Linear Discriminant Analysis and Tone Discrimination Model, Proceedings of IEEE Tencon'02, 30, 7, pp. 959-963, (2002)
  • [4] Li H., Liu J., Liu R., High Performance mandarin Digit String Speech Recognition, ACTA Electronica Sinica, 29, 5, pp. 1-4, (2001)
  • [5] Zuoying W., Jian L., The Speaking rate adaptation algorithm in putonghua continuous speech recognition, ACTA ACUSTICA, 28, 3, pp. 229-234, (2003)
  • [6] Li Z., Yilong L., Cairong Z., Zhenyang W., Application of VQ-HMM to Chinese Spoken Digit Recognition, Journal of Southeast University,, 16, 1, pp. 13-16, (2000)
  • [7] Lu X., Hu W., Wang X., The Mandarin Digit Speech Recognition Based on Artificial Neural Networks, Guangxi Sciences, 11, 4, pp. 320-322, (2004)
  • [8] Babu P., Polur D., Development and optimization of a HMM-ANN hybrid structure for robust recognition of impaired speech signals, (2005)
  • [9] Hong L.Y., Qiao L., Qiang R., Speech Recognition based on Fuzzy Clustering Neural Networks, Chinese Journal of Computers, 29, 10, pp. 1895-1990, (2006)
  • [10] Zhang T., Yanhua, A New Anti-Noise Feature Parameter in Mandarin Digit speech Recognition, Journal of Henan University of Science & Technology(Natural Science), 26, 3, pp. 46-48, (2005)
12