Unscented-Kalman-filtering based approach for single channel speech enhancement

被引：0

作者：

Yang Xi ^{[1
]}

Fan Xiao-ping ^{[1
]}

Luo Xiong ^{[1
]}

机构：

[1] Cent S Univ, Coll Informat Sci & Engn, Res Ctr Automat Engn, Changsha 410075, Peoples R China

来源：

PROCEEDINGS OF 2005 CHINESE CONTROL AND DECISION CONFERENCE, VOLS 1 AND 2 | 2005年

关键词：

speech enhancement; unscented Kalman filter; noise masking; SNR;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

An approach for single channel speech enhancement based on unscented Kalman filtering and masking properties of human auditory system is proposed. In the process of nonlinear time-domain unscented Kalman filtering, the parameter update and the calculation of noise masking are combined. Simulation results show that the proposed method is better than extended Kalman filtering method in SNR and PESQ test.

引用

页码：702 / +

页数：4

共 9 条

[1] A SIGNAL SUBSPACE APPROACH FOR SPEECH ENHANCEMENT
EPHRAIM, Y
VANTREES, HL
[J]. IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, 1995, 3 (04): : 251 - 266
[2] GABREA M, 1999, IEEE SIGNAL PROC MAR, P53
[3] JULIER SJ, 1995, PROCEEDINGS OF THE 1995 AMERICAN CONTROL CONFERENCE, VOLS 1-6, P1628
[4] Klein M, 2002, INT CONF ACOUST SPEE, P537
[5] NELSON AT, 1997, P INT C NEUR NETW HO, P2171
[6] Nishimura R, 2001, IEICE T FUND ELECTR, VE84A, P1027
[7] Paliwal K. K., 1987, Proceedings: ICASSP 87. 1987 International Conference on Acoustics, Speech, and Signal Processing (Cat. No.87CH2396-0), P177
[8] van der Merwe R, 2001, INT CONF ACOUST SPEE, P3461, DOI 10.1109/ICASSP.2001.940586
[9] Wan EA, 2000, ADV NEUR IN, V12, P666

← 1 →