基于小波变换的汉语声调识别新方法

利用小波变换能够模拟人耳听觉特性的优点及其时频局部化特性，对实际语音信号进行分析。根据小波变换模极大值原理以及语音信号的奇异和瞬态特性，提出了一种新的小波变换峰值检测算法。该算法可用于精确提取语音的基音周期，准确进行声调识别。最后，给出应用实例说明了本文算法的有效性。

A New Method For Chinese Tone Recognition Based on Wavelet Transform

Analysis of real speech signals is made on the basis of wavelet transform (WT)which has good time-frequency localization as well as the advantage of being able to simulate the auditory property of ear.With the modulus maximum principle of WT and the regularity and transiency property of speech signals,a peak-value detection algorithm with WT is presented.It has found its applications on exact pitch detection and accurate Chinese tone recognition.Finally,the efficiency of this algorithm isillustrated with an example.