Simon
Godsill
Signal
Processing and Communications (SigProC) Laboratory
Professor of Statistical Signal Processing
Fellow of Corpus Christi College Cambridge
Research
Interests - Signal Inference and its Applications:
- Audio signal processing: source
separation, music analysis and transcription,
noise reduction, audio
restoration, multiple channel audio, sparse/overcomplete
models
- Tracking:
sensor fusion, multiple object tracking, detection, radar, sonar
|
|
|
||||
|
|
|||||
- Signal inference methodololgy: Bayesian methods, Monte Carlo
methods, Markov chain Monte Carlo, Particle Filters (sequential Monte
Carlo), model uncertainty
Research Areas:
Audio and Music Processing (AMP)
The Signal Processing Laboratory
has had long involvement in audio and music processing. Early work in sound
restoration here in the 1980's led to the founding of the successful company
CEDAR Audio Ltd. which produces DSP equipment for remastering
and enhancement of sound in the recording, broadcast and forensic industries.
In current research we are concerned with accurate modelling
of digital audio and automated inference about the parameters and structure of
those models. Research interests include computer
music transcription, audio source separation, musical beat-tracking, chord
recognition, Digital
Audio Restoration, noise reduction, multichannel audio and sparse modelling with overcomplete
dictionaries of atoms. Underpinning much of the work is a Bayesian
statistical modelling approach to audio problems, see
below.
Researchers (current and recent past): Taylan Cemgil, Paul Peeling, Nick
Whitely, Han Lin, Maurice Fallon, Cédric
Févotte, William Fong, Simon
Godsill, Steve Hainsworth, Matt Scarisbrick,
Jaco
Vermaak, Patrick Wolfe, Manuel Davy
Selected papers:
C. Févotte and
S. Godsill. Sparse linear regression in unions of bases
via Bayesian variable selection. IEEE Signal Processing Letters,
13(7):441-444, Jul. 2006.
[ bib
| http ]
P. H. Peeling, C. Li, and
S. J. Godsill. Poisson point process modeling for polyphonic music transcription. Journal
of the Acoustical Society of America Express Letters, 121(4):EL168-EL175,
April 2007. Reused with permission from Paul Peeling, The
Journal of the Acoustical Society of America, 121, EL168 (2007). Copyright
2007, Acoustical Society of America.[ bib
| .html
| .ps.gz | .pdf ]
T. Cemgil, S. J.
Godsill, and C. Fevotte. Variational and Stochastic
Inference for Bayesian Source Separation. Digital Signal Processing,
Accepted for Publication, 2007.
[ bib
| .pdf ]
Févotte
and S.J. Godsill. A Bayesian
approach for blind separation of sparse sources. IEEE
Trans. on Speech and Audio Processing, 2007.
[ bib
]
M. Davy, S.J. Godsill, and J. Idier. Bayesian analysis of polyphonic
western tonal music. Journal of the Acoustical Society of America,
119(4), April 2006.
P.J. Wolfe
and S. J. Godsill. Interpolation
of missing data values for audio signal restoration using a Gabor regression
model. In Proc. IEEE International Conference on
Acoustics, Speech and Signal Processing, March 2005.
C. Févotte
and S. J. Godsill. A
Bayesian approach to time-frequency based blind source separation. In Proc. of IEEE Workshop on Applications of Signal Processing
to Audio and Acoustics, October 2005.
M. Lombardi and S.J. Godsill. On-line
Bayesian estimation of AR signals in symmetric alpha-stable noise. IEEE Trans. on Signal Processing, 2006. [ bib
| .html]
P. J. Wolfe, S. J. Godsill, and W.J. Ng. Bayesian variable
selection and regularisation for time-frequency
surface estimation. Journal of the Royal Statistical Society, Series B 66(3):575-589, 2004. Read paper (with
discussion). [bib | .pdf ]
M.Davy and S. J. Godsill. Bayesian harmonic models for musical signal analysis (with
discussion). In J.M. Bernardo, J.O. Berger, A.P. Dawid,
and A.F.M. Smith, editors, Bayesian Statistics VII. Oxford
University Press, 2003. [ bib
| .ps ]
S. J. Godsill and P. J. W. Rayner. Robust reconstruction and analysis of autoregressive signals in
impulsive noise using the Gibbs sampler. IEEE Trans. on Speech and
Audio Processing, 6(4):352-372, July 1998. [ bib
| .ps.gz ]
Book:
S.J. Godsill and P.J.W. Rayner. Digital
Audio Restoration - a statistical model-based approach (Berlin: Springer-Verlag 1998)
Talks:
Bayesian
harmonic models for musical signal analysis (with M. Davy). Invited lecture
for Seventh Valencia conference on Bayesian Statistics - Tenerife, 2002
Digital Audio Restoration. An introductory talk given at
Helsinki University of Technology, Fall 2003
An
introduction to MCMC methods for sparse overcomplete
sparse audio models. Tutorial for European Union Project HASSIP workshop,
Cambridge, Sept. 2004
Recent Research Projects: HASSIP, MOUMIR, MUSCLE
Tracking
Algorithms
A major challenge in many
application areas is that of detection, classification and tracking of multiple
objects. Classic applications of this include radar and sonar, but the
principles extend into computer vision, robotics and many other areas. We are
aiming to push back the boundaries of current technology where many objects are
present, detection probabilities are low and clutter rates are high. The
methods devised use novel implementations of Monte Carlo Bayesian updating to
carry out joint detection of number, characteristics and position of objects in
cluttered environments.
Researchers (current and
recent past): Sze Kim Pang, Daniel Clark, Jaco
Vermaak, William Ng, Jack Li, Pieter J.P. de Villiers
Selected papers
S.J.
Godsill, J. Vermaak, K-F. Ng, and J-F. Li. Models and algorithms for tracking of manoeuvring objects using variable rate particle filters. Proc.
IEEE, May
2007. [ bib
]
S.J. Godsill. Particle filters for
continuous-time jump models in tracking applications. In ESAIM: PROCEEDINGS
of Oxford Workshop on Particle Filtering, 2007. [ bib
]
J. Vermaak,
N. Ikoma, and S.J. Godsill. Sequential Monte Carlo framework for extended object tracking. IEE Proc.-Radar Sonar Navig.,
152(5):353-363, October 2005.
J. Vermaak,
S. Godsill, and P. Perez. Monte Carlo filtering for multi-target tracking and data
association. IEEE Tr. Aerospace and Electronic Systems,
41(1):309-332, January 2005.
K. Gilholm,
S.J. Godsill, S. Maskell, and D. Salmond.
Poisson models for extended target and group tracking. In Proc. SPIE:
Signal and Data Processing of Small Targets, 2005.
S. J.
Godsill and J. Vermaak. Variable
rate particle filters for tracking applications. In Proc.
IEEE Stat. Sig. Proc., Bordeaux, 2005.
W. Ng, J.F. Li, S.J.
Godsill, and J. Vermaak. A
hybrid approach for online joint detection and tracking for multiple targets.
In IEEE Aerospace Conference, 2005.
S. J.
Godsill and J. Vermaak, Models
and algorithms for tracking using trans-dimensional sequential Monte Carlo.
In Proc. IEEE ICASSP 2004
Research Projects with: DIF-DTC, QinetiQ
Genomic and Life Sciences
Signal Processing
A further topic of great importance is the
interpretation and analysis of genomic data - for example the sequencing of the
human genome. Any improvements achievable in this area are likely to lead to
improvements our understanding of genetics and in
treatment for diseases such as cancer. Work to date has focused on improving
the performance of DNA sequencing machines through very accurate Bayesian
modelling. Currents topics of work include the accurate preprocessing
of microarray data - crucial in identification of the genes active in certain
diseases.
Selected papers:
ˇ Ji Won
Yoon, Simon Godsill, Eriks Kupce,
and Ray Freeman. Deterministic and statistical methods for reconstructing
multidimensional nmr
spectra. Magnetic Resonance in Chemistry, March 2006. [ bib
]
ˇ
N. M. Haan
and S. J. Godsill. Bayesian
models for DNA sequencing. In Proc. IEEE International Conference on
Acoustics, Speech and Signal Processing, 2002
ˇ
N. M. Haan
and S. J. Godsill. A
time-varying model for DNA sequencing data submerged in correlated noise.
In Proc. IEEE Workshop on Statistical Signal Processing, August
2001
ˇ
N.M. Haan and
S.J. Godsill. Sequential methods for DNA sequencing. In Proc. IEEE International Conference on Acoustics, Speech
and Signal Processing, 2001.
ˇ
N. M. Haan
and S. J. Godsill. Modelling electropherogram data for DNA sequencing using variable
dimension MCMC. In Proc. IEEE International
Conference on Acoustics, Speech and Signal Processing, 2000.
Bayesian Computational Methods
for Signal Processing
Underpinning
much of the above applications work is the Bayesian paradigm and associated
algorithms for inference about the parameters and structure of complex systems.
In the Bayesian approach data is combined
with any prior information available in an optimal fashion using probability
distributions. We are particularly concerned with the development of new
methods appropriate to the applications above. These applications are often sequential
in nature (the data arrive one-by-one and a decision/estimate is required with
small or no delay), hence we focus considerable attention on sequential
learning methods such as Sequential Monte Carlo (particle filtering). Other
problems are batch in nature (the data arrive all at once, or we can
wait until all of the data have arrived before processing) - in those cases
batch algorithms can be used, and we focus attention on stochastic simulation
methods such as Markov chain Monte Carlo (MCMC), including those for model
uncertainty problems (reversible jump MCMC, etc.). Novel techniques are developed
to help tailor these methods to the applications at hand.
Projects:
Bayesian inference for partially oberved diffusion
models - EPSRC (in collaboration with Gareth Roberts and Neil Shephard) - researcher Gary
Yang
Selected
papers:
Markov Chain Monte Carlo (MCMC) methods, including model
uncertainty:
ˇ P. J. Wolfe, S. J. Godsill,
and W.J. Ng. Bayesian
variable selection and regularisation for time-frequency surface estimation.
Journal of the Royal Statistical Society, Series B,
2004. Read paper (with discussion).
ˇ
J. Vermaak,
C. Andrieu, A. Doucet,
and S. J. Godsill. Bayesian
model selection of autoregressive processes. J. Time Series Anal.
(In Press).
ˇ
S. J. Godsill. Discussion of `trans-dimensional Markov chain Monte Carlo' by Peter
J. Green. In Highly Structured
Stochastic Systems. OUP, 2003.
ˇ
Doucet, S. J. Godsill, and C. P. Robert. Marginal
maximum a posteriori estimation using MCMC. Statistics
and Computing, 12:77-84, 2002.
ˇ
S. J. Godsill. On the
relationship between Markov chain Monte Carlo methods for model uncertainty.
J. Comp. Graph. Stats.,
10(2):230-248, 2001.
ˇ
Paul T. Troughton
and Simon J. Godsill. MCMC methods for restoration of nonlinearly distorted
autoregressive signals. Signal
Processing, 81(1):83-97, 2001.
ˇ S. J.
Godsill. Inference in symmetric alpha-stable noise using MCMC and the slice
sampler. In Proc.
IEEE International Conference on Acoustics, Speech and Signal Processing,
volume VI, pages 3806-3809, 2000.
ˇ S. J.
Godsill. MCMC and EM-based methods for inference in heavy-tailed processes
with alpha-stable innovations. In Proc. IEEE Signal processing workshop on higher-order
statistics, June 1999. Caesarea, Israel.
ˇ S. J.
Godsill and E. E. Kuruoglu. Bayesian
inference for time series with heavy-tailed symmetric alpha -stable noise
processes. CUED Tech rep INFENG...In Proc.
Applications of heavy tailed distributions in economics, engineering and
statistics, June 1999. Washington DC, USA.
Sequential Monte Carlo (particle filtering)
methods:
O. Cappé,
S.J. Godsill, and E.Moulines. An overview of existing
methods and recent advances in sequential monte carlo. Proc. IEEE, May
2007.[ bib
]
S.J. Godsill, J. Vermaak, K-F. Ng, and J-F. Li. Models and algorithms for tracking of manoeuvring objects using variable rate particle filters. Proc.
IEEE, May 2007.[ bib
]
M. Lombardi and S.J. Godsill.
On-line Bayesian estimation of AR signals in symmetric alpha-stable noise. IEEE Trans. on Signal Processing, 2006. [ bib
| .html]
S. J. Godsill and
J. Vermaak, Models
and algorithms for tracking using trans-dimensional sequential Monte Carlo.
In Proc. IEEE ICASSP 2004
S.J. Godsill and A. Doucet and M.
West. Monte Carlo smoothing for non-linear time series. Journal of the American Statistical
Association. Vol.50, pp. 438-449, 2004
J. Vermaak, S. J. Godsill, and A. Doucet. Radial basis function regression using trans-dimensional sequential
Monte Carlo. In IEEE
Workshop on Statistical Signal Processing, 2003.
J. Vermaak, S. J. Godsill, and A. Doucet. Sequential Bayesian kernel regression.
In Advances in Neural Information Processing Systems
16, Cambridge, MA. MIT Press, 2003.
W. Fong,
S. J. Godsill, A. Doucet, and M. West.
Monte Carlo smoothing with application to speech enhancement. IEEE Trans. on Signal Processing,
50(2):438-449, February 2002.
J. Vermaak, C. Andrieu,
A. Doucet, and S. J. Godsill. Particle methods for Bayesian modelling and enhancement of speech
signals. IEEE Trans. on Speech and
Audio Processing, 10(3):173-185, 2002.
S. J. Godsill and T. C. Clapp. Improvement strategies for Monte Carlo particle filters. In A Doucet,
J. F. G. De Freitas, and N. J. Gordon,
editors, Sequential Monte Carlo Methods in Practice. New York:
Springer-Verlag, 2001.
S. J.
Godsill, A Doucet, and
M West. Maximum
a posteriori
sequence estimation using Monte Carlo particle filters. Ann. Inst.
Stat. Math., 53(1):82-96, March 2001.
Doucet,
S. J. Godsill, and C. Andrieu.
On sequential Monte Carlo sampling methods for Bayesian filtering. Statistics and Computing, 10:197-208, 2000.
T. C. Clapp and S. J. Godsill. Fixed-lag smoothing using sequential importance sampling. In J.M. Bernardo, J.O. Berger, A.P. Dawid,
and A.F.M. Smith, editors, Bayesian Statistics VI, pages 743-752.
Oxford University Press
Tutorial
Materials:
Bayesian
Computer intensive methods for Statistical Signal
Processing (Plenary Address, IEEE Workshop on Statistical Signal
Processing, St Louis August 2003)
On-Line
Bayesian Methods for estimation of non-linear non-Gaussian signals
(Tutorial for Opening Workshop of SAMSI programme, North Carolina, Sept. 2002)
Teaching
Current activities include:
4F7 - Adaptive
Filters and Spectrum Estimation
3F3 - Signal
and Pattern processing
IB Paper 6 -
Signal and Data Analysis
Full list of Publications:
Book - Digital
Audio Restoration - a statistical model-based approach (Springer-Verlag 1998)
Papers (including
downloadable Postscript/pdf)
Bayesian Picture
Gallery
[Cambridge University | CUED |Signal
Processing Group
Updated Sept. 07