人体语音特征提取身份优化验证仿真研究  

Simulation Study on the Identification of Human Voice Sign

作　　者：乔玲玲[1] 郭秀婷[2] 

机构地区：[1]商丘学院计算机工程学院,河南商丘476000 [2]兰州理工大学理学院,甘肃兰州730050

出　　处：《计算机仿真》2017年第2期342-345,共4页Computer Simulation

摘　　要：对人体语音征提取身份优化验证,可为说话人识别奠定基础。进行人体语音征提取身份验证时,应分析人体语音音段韵律特征矢量序列,提取最优音段韵律的高维特征值和特征向量,但是传统方法通过对标注音节的持续采样点数进行分析完成检测,但是不能精确分析人体语音音段韵律特征矢量序列,无法准确提取最优音段韵律的高维特征值和特征向量,存在人体语音征提取身份验证误差大的问题。提出一种改进混沌的人体语音征提取身份优化验证方法。上述方法先融合于混沌理论采集人体发声过程中音段韵律原始信号,将原始韵律信号映射到高维空间实现音段韵律相空间重构,映射相空间中音段韵律间相邻轨道发散的平均变化率,然后利用K-均值聚类的方法对音段韵律的语音帧进行聚类,获取规范化的音段韵律特征矢量序列,将规范化的音段韵律特征矢量序列投影到音段韵律高维核空间中,提取最优音段韵律的高维特征值和特征向量,依据人体语音征提取身份优化验证,仿真结果证明,所提方法特征提取精确度高,能够有效地提升人体语音征提取身份验证的辨识率。An optimization verification method for identity extraction of human voice feature is proposed based on the modified chaos. Firstly,the original signal of segment rhyme during human sound production process is collected integrated with chaos theory, and the original rhyme signal is mapped to high -dimension space to achieve the phase - space reconstruction of segment rhyme. Then, the emanative average change rate of neighbor track among segment rhymes in the phase space is obtained, and the K - mean clustering method is used to cluster the voice frame of segment rhyme. Moreover, the normalized feature vector sequence of segment rhyme is acquired, and the feature vector sequence is projected into the high - dimension nuclear space of segment rhyme. Finally, the optimal high - dimension feature value and feature vector of segment rhyme are extracted, and the verification for identity extraction of human voice feature is optimized. The simulation results show that the method has high feature extraction precision. It can improve the verification recognition rate of identity extraction of human voice feature.

关 键 词：人体发声过程 音段韵律 特征提取 身份验证 

分 类 号：TN912.34[电子电信—通信与信息系统]