Voice activity detector employing generalised Gaussian distribution

被引：27

作者：

Chang, JH ^{[1
]}

Shin, JW

Kim, NS

机构：

[1] Univ Calif Santa Barbara, Dept Elect & Comp Engn, Santa Barbara, CA 93106 USA

[2] Seoul Natl Univ, Sch Elect Engn, Seoul 151742, South Korea

[3] Seoul Natl Univ, INMC, Seoul 151742, South Korea

来源：

ELECTRONICS LETTERS | 2004年 / 40卷 / 24期

关键词：

D O I：

10.1049/el:20047090

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

A novel approach to a voice activity detector (VAD) in noisy environments is presented. The generalised Gaussian distribution (GGD) is employed as a parametric model for noisy speech, which enables tuning to the actual data. According to the experimental results, it was discovered that the proposed GGD model is more effective for the VAD algorithm compared to the conventional Laplacian model.

引用

页码：1561 / 1563

页数：3

共 7 条

[1]

Chang JH, 2001, IEICE T INF SYST, VE84D, P1231

[2] Variability in brain death declaration practices in pediatric head trauma patients [J].

Chang, MY ;

McBride, LA ;

Ferguson, MA .

PEDIATRIC NEUROSURGERY, 2003, 39 (01) :7-9

[3] Speech probability distribution [J].

Gazor, S ;

Zhang, W .

IEEE SIGNAL PROCESSING LETTERS, 2003, 10 (07) :204-207

[4]

Mallat S., 1989, IEEE PAMI, V11, P674

[5] DISTRIBUTION SHAPE OF 2-DIMENSIONAL DCT COEFFICIENTS OF NATURAL IMAGES [J].

MULLER, F .

ELECTRONICS LETTERS, 1993, 29 (22) :1935-1936

[6]

REININGER RC, 1983, IEEE T COMMUN, V31

[7] A statistical model-based voice activity detection [J].

Sohn, J ;

Kim, NS ;

Sung, W .

IEEE SIGNAL PROCESSING LETTERS, 1999, 6 (01) :1-3

← 1 →