基于DTW算法对婴幼儿语音的分析

Analysis of infant speech based on DTW algorithm

针对婴幼儿语音分析问题,通过基音频率识别、端点技术检测、DTW算法等原理,运用MATLAB软件,识别出搜集的声音片段的性别,并对声音进行谱曲,进行误差分析。通过查阅相关资料在鉴别男女生声音的模型基础上建立出婴幼儿的性别识别模型,并对模型进行误差检验,使得模型的拟合度大大提高。

In order to analyze the voice of infants and young children,MATLAB software is used to identify the gender of the collected sound fragment through the principles of basic sound frequency recognition,endpoint technology detection,and DTW algorithm,and to compose the sound and analyze the error.The gender recognition model of infants and young children was established based on the model of identifying male and female voices by consulting relevant data,and the error test of the model was carried out,so that the similarity of the model was greatly improved.