School of Electronic, Electrical and Systems Engineering

This project has focused upon the machine cocktail party problem. A new multi-modal approach which utilizes both audio and video measurements has been proposed for separating both static and moving speakers. The video information provides information about the location and velocity of the sources. The separation is performed either with beamforming or independent component analysis dependent upon the velocity of the source. The location information is used either to steer the beamformer or intelligently initialize the fast ICA algorithm.

Theses

• A. Aubrey, ''Exploiting the bimodality of speech in the cocktail party problem'', PhD thesis
• S. M. Naqvi , ''Multimodal methods for blind source separation of audio sources'', PhD thesis

Journal Papers

• Y. Luo; W. Wang; Chambers, J.A.; Lambotharan, S.; Proudler, I., "Exploitation of source nonstationarity in underdetermined blind source separation with advanced clustering techniques," Signal Processing, IEEE Transactions on , vol.54, no.6, pp. 2198-2212, June 2006
• Zhang, Y.; Chambers, J.A., "Variable tap-length natural gradient blind deconvolution=equalisation algorithm," Electronics Letters , vol.43, no.14, pp.-, July 5 2007.
• Y. Zhang; N. Li; Chambers, J.A.; Sayed, A.H., "Steady-State Performance Analysis of a Variable Tap-Length LMS Algorithm," Signal Processing, IEEE Transactions on , vol.56, no.2, pp.839-845, Feb. 2008.
• Y. Zhang, N. Li, J. A. Chambers and Y. Hao, “New gradient based variable step-size LMS algorithms”. EURASIP Journal on Advances in Signal Processing, Vol.8 (2), 1-9, volume 2008.
• N. Li, Y. Zhang, Y. Hao, J. A. Chambers, A new variable step-size NLMS algorithm designed for applications with exponential decay impulse responses, Signal Processing, Volume 88, Issue 9, Pages 2346-2349, September 2008.
• S. M. Naqvi, Y. Zhang, M. Yu, and J. A. Chambers, “A multimodal approach to blind source separation of moving sources,” submitted after review to IEEE Transactions on Multimedia.
• A. Aubrey, Y. Hicks, and J. Chambers, “Visual Voice Activity Detection with Optical Flow,” In preparation for submission to IET Proceedings on Signal Processing.

Conference Papers

• Aubrey, A.; Hicks, Y.; Sanei, S.; Chambers, J., "Study of Video Assisted BSS for Convolutive Mixtures," Digital Signal Processing Workshop, 12th - Signal Processing Education Workshop, 4th , vol., no., pp.273-277, 24-27 Sept. 2006.
• A. Aubrey, J. Lees, Y. Hicks, and J. Chambers, “Using the Bimodality of Speech for Convolutive Frequency Domain Blind Source Separation,” in IMA 7th International Conference on Mathematics in Signal Processing. December, 2006.
• A. Aubrey, B. Rivet, Y. Hicks, L. Girin, J. Chambers, and C. Jutten, “Two Novel Visual Voice Activity Detectors Based on Appearance Models and Retinal Filtering,” In 15th European Signal Processing Conference (EUSIPCO). Poland, 2007.
• B. Rivet, A. Aubrey, L. Girin, Y. Hicks, C. Jutten, and J. Chambers, “Development and comparison of two approaches for visual speech analysis with application to voice activity detection,” Int. Conference On Auditory Visual Speech Processing (AVSP). The Netherlands, 2007.
• S. M. Naqvi, Y. Zhang and J. A. Chambers, “Evaluation of emerging frequency domain convolutive blind source separation algorighms based on real recordings,” The 5th IEEE Sensor Array and Multichannel Signal Processing Workshop, July, Germany, 2008.
• S. M. Naqvi, Y. Zhang and J. A. Chambers, “A multimodal approach for frequency domain blind source separation for moving sources in a room,” 1st IARP workshop on Cognitive information processing, Greece, 2008.
• S. M. Naqvi, Y. Zhang, T. Tsalaile, S. Sanei and J. A. Chambers, “A multimodal approach for frequency domain independent component analysis with geometrically-based initialization,” EUSIPCO 2008, Switzerland.
• Sanei, S.; Naqvi, S.M.; Chambers, J.A.; Hicks, Y., "A Geometrically Constrained Multimodal Approach for Convolutive Blind Source Separation," Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on , vol.3, no., pp.III-969-III-972, 15-20 April 2007.
• Y. Zhang, N. Li, and J. A. Chambers, “A new gradient based variable step-size LMS algorithm”, 8th International Conference on Signal Processing, Guilin, China, 2006.
• L. L. Li, Y. Zhang and J. A. Chambers, “Variable length adaptive filtering within incremental learning algorithm for distributed networks,” Asilomar Conference Signals, Systems and computers, Pacific Grove, CA.Oct. 2008.
• L. L. Li, Y. Zhang and J. A. Chambers, “Steady-state performance of incremental learning over distributed networks for non-Gaussian data,”  9th International Conference on Signal Processing, Beijing, China, Oct. 2008.
• S. M. Naqvi, Y. Zhang and J. A. Chambers, “Multimodal blind source separation for moving sources”, presented at ICASSP 2009, Taipei, April 2009.