论文部分内容阅读
详细分析了目前语音识别系统中普遍采用的Mel频率倒谱系数(MFCC)特征参数的提取过程和动态时间规整(DTW)识别算法流程,提出了一种在NiosII SOPC软核平台上通过提取和分析语音信号的MFCC特征参数实现语音识别解决方案。
This paper analyzes in detail the extraction process of Mel Frequency Cepstral Coefficients (MFCC) feature parameters and the flow of dynamic time warping (DTW) recognition algorithm commonly used in speech recognition system, and proposes a method to extract and analyze the parameters of NiosII SOPC soft core platform, Speech signal MFCC feature parameters to achieve voice recognition solutions.