Wavelet Analysis for Robust Speech Processing and Applications

Robustness to acoustic background noise is a highly challenging field in modern speech communications. A state-of-the-art research of the techniques has been motivated by the need for efficient speech feature extraction methods, the need for enhancing the quality of the recorded speech signal corrupted by environmental noise, and the need for improving recognition performance in harsh environments. This book, therefore, provides novel wavelet methods which are developed for addressed topics. The book begins with the inter-pretation of wavelets advantages for speech processing. The next chapters present phonetic and speech classifiers which are designed based on reliable wavelet features and by applying adaptive-threshold based and machine learning based approaches. Afterwards, based on the adaptive noise estimate of the so-called quantile filtering tech-nique and the optimal wavelet shrinkage, speech enhancement algo-rithms are developed for hearing aids and for enhancing signal quality in order to increase speech recognition performance. The book is intended for advanced undergraduate and graduate students, as well as for scientific researchers in the field of interest.

Author: Tuan Van Pham

Publisher: VDM Verlag Dr. Mueller E.K.

Publication Date: May 26, 2008

Number of Pages: 176 pages

Binding: Paperback or Softback

ISBN-10: 3639024168

ISBN-13: 9783639024166