Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (35): 172-174.

• 工程与应用 • Previous Articles     Next Articles

Application of fractal and neural network in UV decision

CHEN Jia1,TAN Guan-zheng1,YE Ji-xiang1,2   

  1. 1.College of Information Science and Engineering,Central South University,Changsha 410083,China
    2.Changsha University of Science & Technology,Changsha 410076,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-12-11 Published:2007-12-11
  • Contact: CHEN Jia

分形理论和神经网络在清浊音判决中的应用

陈 佳1,谭冠政1,叶吉祥1,2   

  1. 1.中南大学 信息科学与工程学院,长沙 410083
    2.长沙理工大学,长沙 410076
  • 通讯作者: 陈 佳

Abstract: we can take the speech signal as a kind of chaos signal and adopt the approach of chaos signal process to analysis the speech signal.The fractal can indicate how chaos the signal is.And because of the different mechanic of pronounce of the unvoiced and the voiced,they must be different in the fractal value.We calculate the mean of fractal value to identify the voiced and voiceless.then present the DP parameter.taking the fractal and DP as a vector,recognise the signal by the ANN.

Key words: fractal, surd, sonant, ANN

摘要: 采用混沌信号处理方法中的分形理论对信号进行分析。分形维数很好的体现了信号的混沌程度,而清音和浊音由于在发声原理上的不同,清音类似于噪声,浊音具有近似的周期性,在分形维这个特征上体现出差异。首先对语音信号分帧求分形维轨迹,计算出平均分形维,然后在分形维参数的基础上提出DP值特征参数,以分形维与DP值作为一个特征向量,采用BP神经网络进行识别,得到了很好的识别效果。

关键词: 分形维, 清音, 浊音, 人工神经网络