摘要
以相空间重构理论为基础,采用Takens定理重构语音信号相空间并提取相似序列重复度(RPT)特征参数。利用清浊音RPT参数的差异,提出并实现了一种采用BP神经网络进行非线性清浊音判决的方法,得到了明显优于传统算法的结果。
Based on Takens theory, time delay method is used to reconstruct phase space of speech in this paper. In hyper dimensional phase space, similar sequence repeatability (RPT) of speech are calculated. At the same time, according to the difference of RPT between voice and unvoice, BP neural network is applied to nonlinear V/UV decision, and the results are better than other tranditional methods. The method proposed in this paper provides a new way for feature extraction and speech recognition.
出处
《通信学报》
EI
CSCD
北大核心
2003年第6期16-22,共7页
Journal on Communications
