Dual-channel speech separation using interaural time difference with Generalized Gaussian Mixture Model

DOI

10.2991/icitmi-15.2015.194How to use a DOI?

Keywords

interaural time difference (ITD) statistics, Generalized Gaussian Mixture Model, correlation coefficient, time-frequency mask

Abstract

In this letter we present a novel speech separation scheme using two microphones. The proposed method utilizes the estimation of interaural time difference (ITD) statistics for the separation of mixed speech sources. The novelties of this paper consist in the use of Generalized Gaussian Mixture Model (GGMM) for speech separation frame by frame and cross-correlation coefficient for distributed parameter selection. The proposed model can be extended to audio enhancement. Our objective quality evaluation experiments demonstrate the effectiveness of the proposed methods and show significant quality improvements over the conventional dual ITD based methods.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhaogui Ding
AU  - Liming Zhang
AU  - Longbiao Wang
AU  - Weifeng Li
PY  - 2015/10
DA  - 2015/10
TI  - Dual-channel speech separation using interaural time difference with Generalized Gaussian Mixture Model
BT  - Proceedings of the 4th International Conference on Information Technology and Management Innovation
PB  - Atlantis Press
SP  - 1157
EP  - 1163
SN  - 2352-538X
UR  - https://doi.org/10.2991/icitmi-15.2015.194
DO  - 10.2991/icitmi-15.2015.194
ID  - Ding2015/10
ER  -