Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/67517
Title: Improved expectation-maximization framework for speech enhancement based on iterative noise estimation
Authors: Li, T
Lun, DPK 
Shen, TW
Keywords: nNise power spectral density estimation
Speech enhancement
EM algorithm
Iterative regularization
Issue Date: 2015
Publisher: Institute of Electrical and Electronics Engineers
Source: 2015 IEEE International Conference on Digital Signal Processing (DSP), Singapore, 21-24 July 2015, p.287-291 How to cite?
Abstract: Recently, our team developed a novel Expectation Maximization (EM) framework for speech enhancement. It gives a significantly improved estimation of the speech power spectrum that outperforms many traditional approaches. In this paper, we further extend the EM framework by including an efficient iterative noise estimation algorithm, which improves the estimation of the noise power spectrum from the noisy observation. Besides, we notice that some speech frames, particularly those with high signal to noise ratio (SNR), need to be monitored closely during the iterative enhancement process, or spectral distortion may result. A stopping criterion is thus developed to stop the iteration when a good result has been achieved. Experimental results show that the new approach gives a significant improvement over the original EM framework and also traditional speech enhancement methods.
URI: http://hdl.handle.net/10397/67517
ISBN: 978-1-4799-8058-1 (electronic)
978-1-4799-8057-4 (USB)
978-1-4799-8059-8 (print on demand(PoD))
ISSN: 1546-1874
EISSN: 2165-3577
DOI: 10.1109/ICDSP.2015.7251877
Appears in Collections:Conference Paper

SFX Query Show full item record

Page view(s)

16
Last Week
2
Last month
Citations as of Dec 17, 2017

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.