期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

基于长时性特征的音位属性检测方法

Phonological Attribute Detection Method Based on Long-term Features
在线阅读 下载PDF
导出
摘要 提出一种基于长时性信息的音位属性检测方法,该方法通过高、低两层时间延迟神经网络(TDNN)进行实现,低层TDNN在短时特征上进行音位属性的检测,高层TDNN在低层检测结果的基础上,对更长时段上的信息进行融合。实验结果表明,引入长时性特征使得音位属性检测率提升约3%,将音位属性后验概率作为音素识别系统的观测特征,使用长时性特征的识别结果提升约1.7%。 A novel phonological attribute detection method based on long-term information is presented.This method is comprised of high-level and low-level Time-delayed Neural Networks（TDNN）.The low-level TDNN carries out phonological attribute detection on the basis of short-term features,and the high-level TDNN is based on the low-level output and considering the long-term information,and fully taps the relation between speech signals in time.Experimental results show that,compared by the detection using short-term features,the introduction of phonological attribute based on long-term features improves detection rate with 3%.In addition,this paper puts the phonological attribute in phoneme recognition experiments,the results improveing 1.7% in Hidden Markov Model（HMM）-based speech recognition system.
作者 许友亮 张连海 屈丹 牛铜
机构地区 解放军信息工程大学信息工程学院
出处 《计算机工程》 CAS CSCD 2012年第11期160-162,166,共4页 Computer Engineering
基金 国家自然科学基金资助项目(61175017)
关键词 音位属性 长时特征 层级结构 人工神经网络 隐马尔可夫模型 音素识别 phonological attribute long-term features hierarchical structure Artificial Neural Network（ANN） Hidden Markov Model（HMM） phoneme classification
分类号 TP391.4 [自动化与计算机技术—计算机应用技术]
  • 相关文献

参考文献10

  • 1Dusan S, Rabiner L R. On Integrating Insights from Human Speech Perception into Automatic Speech Recognition[C]//Proc. of Conference on International Speech Communication Association. Lisbon, Portugal: [s. n.], 2005: 1233-1236.
  • 2Chen I F, Wang Hsin-Min. An Investigation of Phonological Feature Systems Used in Detection-based ASR[C]//Proc. of Conference on Chinese Spoken Language Processing. Kunming, China: [s. n.], 2008: 1-4.
  • 3King S, Taylor P. Detection of Phonological Features in Con- tinuous Speech Recognition Using Neural Networks[J]. Computer Speech and Language, 2000, 14(4): 333-353.
  • 4Rajamanohar M, Fosler L E. An Evaluation of Hierarchical Articulatory Feature Detectors[C]//Proc. of IEEE Workshop on Automatic Speech Recognition and Understanding. San Juan, Puerto Rico: IEEE Press, 2005: 349-354.
  • 5Chen I F, Wang Hsin-Min. Articulatory Feature Asynchrony Analysis and Compensation in Detection-based ASR[C]//Proc. of Conference on International Speech Communication Association. Lisbon, Portugal: [s. n.], 2009: 3059-3062.
  • 6Chen B, Zhu Qifeng, Morgan N. Learning Long-term Temporal Features in LVCSR Using Neural Networks[C]//Proc. of Conference on Spoken Language Processing. Jeju, Korea: [s. n.], 2005: 1233-1236.
  • 7李晨冲,董滨,潘复平,曾兴雯,颜永红.汉语普通话易混淆音素的识别[J].计算机工程,2009,35(23):201-203. 被引量：4
  • 8Ganapathy S, Thomas S, Hermansky H. Comparison of Modul- ation Features for Phoneme Recognition[C]//Proc. of IEEE International Conference on Acoustics Speech and Signal Processing. Dallas, USA: IEEE Press, 2010: 5038-504,1.
  • 9Ketabdar H, Bourlard H. Hierarchical Integration of Phonetic and Lexical Knowledge in Phone Posterior Estimation[C]//Proc. of IEEE International Conference on Acoustics Speech and Signal Processing. Las Vegas, USA: IEEE Press, 2008: 4065-4068.
  • 10Le V B, Lamel L, Gauvain J L. Multi-style MLP Features for BN Transcription[C]//Proc. of IEEE International Conference on Acoustics Speech and Signal Processing. Dallas, USA: IEEE Press 2010: 4866-4869.

二级参考文献1

共引文献3

计算机工程
2012年 第11期

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部