文件名称:10.1.1.4.2135
下载
别用迅雷、360浏览器下载。
如迅雷强制弹出，可右键点击选“另存为”。
失败请重下，重下不扣分。
如迅雷强制弹出，可右键点击选“另存为”。
失败请重下，重下不扣分。
介绍说明－－下载内容均来自于网络，请自行研究使用
Usually, speaker recognition systems do not take into account the short–term dependence between the vocal source and the vocal tract. A feasibility study that retains this dependence is presented here. A model of joint probability functions of the pitch and the feature vectors is proposed. Three strategies are designed and compared for all female speakers taken from the SPIDRE corpus.
The fi rst operates on all voiced and unvoiced speech segments (baseline strategy). The second strategy considers only the voiced speech segments and the last includes the short–term pitch information along with the standard MFCC. We use two pattern recognizers: LVQ–SLP and GMM. In all cases, we observe an increase in the identifi cation rates and more specifi cally when using a time duration of 500 ms (6 higher).-Usually, speaker recognition systems do not take into account the short–term dependence between the vocal source and the vocal tract. A feasibility study that retains this dependence is presented here. A model of joint probability functions of the pitch and the feature vectors is proposed. Three strategies are designed and compared for all female speakers taken from the SPIDRE corpus.
The fi rst operates on all voiced and unvoiced speech segments (baseline strategy). The second strategy considers only the voiced speech segments and the last includes the short–term pitch information along with the standard MFCC. We use two pattern recognizers: LVQ–SLP and GMM. In all cases, we observe an increase in the identifi cation rates and more specifi cally when using a time duration of 500 ms (6 higher).
The fi rst operates on all voiced and unvoiced speech segments (baseline strategy). The second strategy considers only the voiced speech segments and the last includes the short–term pitch information along with the standard MFCC. We use two pattern recognizers: LVQ–SLP and GMM. In all cases, we observe an increase in the identifi cation rates and more specifi cally when using a time duration of 500 ms (6 higher).-Usually, speaker recognition systems do not take into account the short–term dependence between the vocal source and the vocal tract. A feasibility study that retains this dependence is presented here. A model of joint probability functions of the pitch and the feature vectors is proposed. Three strategies are designed and compared for all female speakers taken from the SPIDRE corpus.
The fi rst operates on all voiced and unvoiced speech segments (baseline strategy). The second strategy considers only the voiced speech segments and the last includes the short–term pitch information along with the standard MFCC. We use two pattern recognizers: LVQ–SLP and GMM. In all cases, we observe an increase in the identifi cation rates and more specifi cally when using a time duration of 500 ms (6 higher).
相关搜索： LVQ
(系统自动生成,下载前可以参看下载内容)
下载文件列表
10.1.1.4.2135.pdf