Multiple moving speaker tracking via degenerate unmixing estimation technique and Cardinality Balanced Multi-target Multi-Bernoulli Filter (DUET-CBMeMBer)

Chong, N. and Wong, S. and Vo, B.T. and Nordholm, S. and Murray, I. 2014. Multiple moving speaker tracking via degenerate unmixing estimation technique and Cardinality Balanced Multi-target Multi-Bernoulli Filter (DUET-CBMeMBer), in IEEE Ninth International Conference on Intelligent Sensors, Sensor Networks and Information Processing, Apr 21 2014. Singapore: IEEE.

Source Title

Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP),

Source Conference

IEEE Ninth International Conference on Intelligent Sensors, Sensor Networks and Information Processing

DOI

10.1109/ISSNIP.2014.6827608

ISBN

978-1-4799-2842-2

School

Department of Electrical and Computer Engineering

URI

http://hdl.handle.net/20.500.11937/24543

Collection

Curtin Research Publications

Abstract

The "cocktail party problem" has always been a challenging problem to solve and many blind source separation algorithms have been proposed as solutions. This problem has mainly been discussed for non-moving sound sources but it still remains for moving sound sources and high acoustic reverberations. The ability to localise and track multiple moving speakers is a pre-requisite to solving this problem. The aim of this paper is to show that a combination of Degenerate Unmixing Estimation Technique and a Cardinality Balanced Multitarget Multi-Bernoulli Filter provides a viable way to track multiple sound sources and subsequently address the problem of sound separation for moving targets.