An Efficient Activation Function for Blind Separation of Speech Signals

DOI

10.2991/icaicte.2013.82How to use a DOI?

Keywords

Blind Separation, Speech Signals, Activation Function.

Abstract

The aim of this paper is to study blind separation of speech signals based on feedforward neural network with sine ac-tivation function. The natural gradient is a major contribution to blind source separation (BSS) because its performance is independent of the mixing matrix. Computer simulations are provided by speech signals as sources (ten speakers) and in the instantaneous mixing case show that the algorithm with sine activation function converges faster and recovers independent sources better than other activation functions.

Copyright

© 2013, 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  - Shouyu Sun
PY  - 2013/08
DA  - 2013/08
TI  - An Efficient Activation Function for Blind Separation of Speech Signals
BT  - Proceedings of the 2013 International Conference on Advanced ICT and Education
PB  - Atlantis Press
SP  - 393
EP  - 396
SN  - 1951-6851
UR  - https://doi.org/10.2991/icaicte.2013.82
DO  - 10.2991/icaicte.2013.82
ID  - Sun2013/08
ER  -