Imperial College London > Talks@ee.imperial > COMMSP Seminar > A Multimodal Approach to Blind Source Separation of Moving Sources

A Multimodal Approach to Blind Source Separation of Moving Sources

Add to your list(s) Download to your calendar using vCal

  • UserProf. Jonathon Chambers, Loughborough University
  • ClockWednesday 01 July 2009, 14:00-15:00
  • HouseEE Dept. Room 403A.

If you have a question about this talk, please contact Danilo Mandic.

The late Professor Colin Cherry, Department of Electrical and Electronic Engineering, Imperial College, is famous for describing the cocktail party problem in work published in 1953. This problem corresponds to mimicing within a machine the ability of a human to follow one conversation whilst others are on going. The work presented in this talk will therefore be a step towards solving the machine cocktail party problem and leveraging findings within pyschology that human auditory perception is commonly multimodal, i.e. visual cues are fundamental to the process.

In particular, the problem of blind source separation (BSS) of moving sources will be introduced. The challenge of BSS for moving sources is that the mixing filters are time-varying; thus the unmixing filters should also be time-varying, however these are difficult to calculate in real time. In the proposed approach, the visual modality is utilized to facilitate the separation for both stationary and moving sources. The movement of the sources is detected by a 3-D tracker based on video cameras. Positions and velocities of the sources are obtained from the 3-D tracker based on a Markov Chain Monte Carlo particle filter, which results in high sampling efficiency. The full BSS solution is formed by integrating a frequency-domain BSS algorithm and beamforming. Experimental results confirm that by utilizing the visual modality, the proposed algorithm not only improves the performance of the BSS algorithm and mitigates the permutation problem for stationary sources, but also provides an aceptable performance for moving sources in a low reverberant enviroment.

Speaker: Jonathon Chambers is Professor of Communications and Signal Processing within the Department of Electronic and Electrical Engineering at Loughborough University, he also leads the Advanced Signal Processing Group and is the Deputy Head of Department, Research. He has served as an Associate Editor for IEEE Trans. for more than ten years and currently serves on the IEEE Signal Processing Theory and Methods Technical Committee and is the Technical Programme Chair for the IEEE Workshop on Statistical Signal Procesing 2009, Cardiff, and the Technical Programme Co-Chair for the flagship conference of the IEEE Signal Processing Society ICASSP 2011 , Prague. He has graduated approaching 50 PhD students and authored/co-authored two research monographs and more than 300 journal and conference works. He worked within the Department of Electrical and Electronic Engineering at Imperial College between 1994 and 2001.

This talk is part of the COMMSP Seminar series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


Changes to Talks@imperial | Privacy and Publicity