北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 1999, Vol. 22 ›› Issue (1): 15-19.

• 分层与分割概念及其在ATM网管系统中的应用 • 上一篇    下一篇

基于IIR两通道双正交滤波器组的小波包分析

闫桂霞, 赵尔沅   

  1. 北京邮电大学电信工程学院, 北京 100876
  • 收稿日期:1998-01-16 出版日期:1999-01-10
  • 基金资助:
    国家自然科学基金资助项目(69572008)

Wavelet Packets Analysis Based on IIR Two-Channel#br# Biorthogonal Filter Banks

Yan Guixia Zhao Eryuan   

  1. School of Telecommunication Engineering, Beijing University ofPosts and Telecommunications, Beijing 100876
  • Received:1998-01-16 Online:1999-01-10

摘要: 使用一种IIR型两通道双正交滤波器组结构, 分析-合成部分的设计仅归结为一个传输函数的设计.据此, 在IIR最大平坦滤波器组的情况下构造了双正交小波基.将其用于实现小波包分析, 采用信息熵函数作信息代价函数, 动态调整最佳基, 使得频谱分段最适合信号特征.对大量音频信号进行处理, 与使用同阶的Daubechies小波的小波变换以及小波包变换的处理结果相比, 性能有非常明显的提高.

关键词: 数字信号处理, 两通道双正交滤波器组, 小波包变换

Abstract: A framework of IIR two-channel biorthogonal filter banks is used here. The design of the analysis/synthesis systems reduces to the design of a single transferfunction. The wavelet bases corresponding to the IIR maximally flat filters aregenerated. Use the wavelet bases in wavelet packets analysis, take information entropy function as information cost function, adjust the best bases dynamically,and the section of signal's spectrum fits its characteristic well. A lot of audio signals are processed. Compared to that of the wavelet transform and the wavelet packet transform using the Daubechies wavelet which has the same order withour IIR wavelet, our properties are better.

Key words: digital signal processing, two-channel biorthogonal filter banks, wavelet packet transform

中图分类号: