Speaker recognition and speaker normalization by projection to speaker subspace
Speaker recognition and speaker normalization by projection to speaker subspace
复制标题
通过投影到说话人子空间进行说话人识别和说话人归一化
DOI：
10.1109/icassp.1996.541096
复制
发表时间：
1996
期刊：
影响因子：
--
通讯作者：
Masayuki Nishijima
中科院分区：
文献类型：
--
作者：
Y. Ariki;S. Tagashira;Masayuki Nishijima
一个说话人被认为有他自己的子空间，其中包括他的语音信息。然而，传统的说话人无关的HSPER忽略了说话人子空间，并收集广泛分布在观察空间中的语音数据。然后，他们造成的概率分布平坦的障碍和由此产生的识别错误。为了解决这个问题，我们提出了一种方法（1）分离说话人的特征，通过构建个人的说话人子空间，（2）识别说话人的基础上的子空间和（3）产生说话人归一化的语音数据投影到他的子空间，并识别他们。
An individual speaker is thought to have his own subspace in which his phonetic information is included. However, conventional speaker-independent HMMs ignore the speaker subspaces and gather speech data spread widely in the observation space. Then they cause probability distribution flatness of HMMs and the resultant recognition errors. To solve this problem, we propose a method (1) to separate the speaker characteristics by constructing the individual speaker subspace, (2) to recognize speakers based on the subspaces and (3) to produce speaker normalized speech data by projecting speech data into his subspace and to recognize them.