现代电子技术2018，Vol.41Issue(10)：169-171,3.DOI:10.16652/j.issn.1004⁃373x.2018.10.043
基于子波方差和尺度波长关系的语音识别算法研究
Speech recognition algorithm based on relationship between wavelet variance and scale wavelength
陈理 1杨雪梅 1毕春艳 1陈莹莹1
作者信息
- 1. 四川大学锦江学院,四川 眉山620860
- 折叠
摘要
Abstract
To improve the precision of speech recognition,an improved pitch detection algorithm for applying the wavelet variance analysis means to the pitch detection is proposed. In the algorithm,Daubechies wavelet is selected to an decompose sig-nals to an appropriate scale,and Sombrero wavelet is then used to perform wavelet variance analysis. The numerical value of pitch period can be accurately located based on the relationship between wavelet scale and its wavelength. The algorithm was ap-plied to the detection of Chinese tones. The simulation results show that the algorithm can adjust the precision and range of the scan scale according to actual demands,and has adjustability.关键词
子波方差/尺度波长/子波能量/基音检测/声调识别/语音识别算法Key words
wavelet variance/scale wavelength/wavelet energy/pitch detection/tone recognition/speech recognition algo-rithm分类
信息技术与安全科学引用本文复制引用
陈理,杨雪梅,毕春艳,陈莹莹..基于子波方差和尺度波长关系的语音识别算法研究[J].现代电子技术,2018,41(10):169-171,3.
