Speaker segmentation based on discriminative deep belief networks - 文章详情页

作者：

Ma, Yong (Ma, Yong.) | Bao, Changchun (Bao, Changchun.) (学者：鲍长春) | Xia, Bingyin (Xia, Bingyin.)

收录：

EI Scopus PKU CSCD

摘要：

A　discriminative　deep　belief　network　(DDBN)　based　on　the　Fisher　criterion　is　used　here　to　calculate　the　super-vector　feature　space　of　speech　signals.　The　network　extracts　the　feature　codebook　of　the　speaker　that　is　superior　to　the　one　from　the　traditional　deep　belief　network　(DBN)　algorithm　for　multi-speaker　clustering　and　segmentation.　Evaluations　on　the　multi-speaker　audio　stream　corpus　generated　from　the　TIMIT　database　show　that　the　speaker　segmentation　algorithm　based　on　the　DDBN　with　the　Fisher　criterion　performs　better　than　the　traditional　Bayesian　information　criterion　(BIC)　method　and　the　DBN　method.

关键词：

Vector spaces Clustering algorithms Bayesian networks

作者机构：

[ 1 ] [Ma, Yong]School of Electronic Information and Control Engineering, Beijing University of Technology, Beijing 100124, China
[ 2 ] [Ma, Yong]School of Physics and Electronic Engineering, Jiangsu Normal University, Xuzhou 221009, China
[ 3 ] [Bao, Changchun]School of Physics and Electronic Engineering, Jiangsu Normal University, Xuzhou 221009, China
[ 4 ] [Xia, Bingyin]School of Electronic Information and Control Engineering, Beijing University of Technology, Beijing 100124, China

通讯作者信息：

电子邮件地址：

baochch@bjut.edu.cn

查看成果更多字段

成果类型
所属机构

所有年份指定年份从至