计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (3): 120-121.

• 数据库、信号与信息处理 • 上一篇    下一篇

小波消噪和ICA在语音信号分离中的应用

王晓伟1,石林锁1,杨 隆2,鲁 秘3   

  1. 1.第二炮兵工程大学 502教研室,西安 710025
    2.第二炮兵工程技术总队,河南 洛阳 471031
    3.中国人民解放军 91515部队
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2012-01-21 发布日期:2012-01-21

Application of wavelet denoising and ICA in separating speech signals

WANG Xiaowei1, SHI Linsuo1, YANG Long2, LU Mi3   

  1. 1.No.502 Faculty, The Second Artillery Engineering University, Xi’an 710025, China
    2.Chief Engineering Technology Troop of Second Artillery, Luoyang, Henan 471031, China
    3.No.91515 Unit of PLA
  • Received:1900-01-01 Revised:1900-01-01 Online:2012-01-21 Published:2012-01-21

摘要: 为了消除语音信号分离中仍存在的部分混叠声音,提出一种基于小波消噪和独立分量分析(ICA)结合的信号分离方法。该方法将小波变换和独立分量分析结合,利用小波变换的去噪作用,滤除原始语音信号中的噪声后作为ICA的输入信号,采用FastICA算法在小波域进行独立分量分析,对输入信号实施分离。实验结果表明,该方法大大调高了传统独立分量分析对语音信号的分离效果。

关键词: 独立分量分析, 小波消噪, 语音分离

Abstract: A separation method is presented to remove the existence of part supposed speech based on wavelet transform and Independent Component Analysis(ICA). Wavelet transform can remove the process noise in the original, from which the de-noise signals come out and it can be used as the input of ICA. FastICA is used to separate the input signals in wavelet domain to gain the low noise independent speech signal. The experiment results show that the proposed method improves the separation effects of the conventional method.

Key words: Independent Component Analysis(ICA), wavelet de-noising, speech signal separation