Title
Source counting in speech mixtures using a variational EM approach for complex WATSON mixture models
Abstract
In this contribution we derive a variational EM (VEM) algorithm for model selection in complex Watson mixture models, which have been recently proposed as a model of the distribution of normalized microphone array signals in the short-time Fourier transform domain. The VEM algorithm is applied to count the number of active sources in a speech mixture by iteratively estimating the mode vectors of the Watson distributions and suppressing the signals from the corresponding directions. A key theoretical contribution is the derivation of the MMSE estimate of a quadratic form involving the mode vector of the Watson distribution. The experimental results demonstrate the effectiveness of the source counting approach at moderately low SNR. It is further shown that the VEM algorithm is more robust with respect to used threshold values.
Year
DOI
Venue
2014
10.1109/ICASSP.2014.6854924
Acoustics, Speech and Signal Processing
Keywords
Field
DocType
Fourier transforms,array signal processing,blind source separation,expectation-maximisation algorithm,least mean squares methods,microphone arrays,speech processing,statistical distributions,MMSE estimate,VEM algorithm,Watson distributions,active sources,blind source separation,complex WATSON mixture models,low SNR,mode vectors,normalized microphone array signal distribution,short-time Fourier transform domain,signal suppression,source counting approach,speech mixtures,variational EM approach,variational expectation maximization algorithm,Bayes methods,Blind source separation,Directional statistics,Number of speakers
Normalization (statistics),Pattern recognition,Quadratic form,Mode (statistics),Model selection,Microphone array,Fourier transform,Artificial intelligence,Mathematics,Mixture model
Conference
ISSN
Citations 
PageRank 
1520-6149
5
0.54
References 
Authors
5
4
Name
Order
Citations
PageRank
Lukas Drude19511.10
Aleksej Chinaev2223.05
Dang Hai Tran Vu3444.01
Reinhold Haeb-Umbach41487211.71