Research Interests
Nonstationary signal processing
Understanding and utilising nonstationarity to produce superior results both in existing problems solved using the stationarity assumption, and previously intractable problems.

Acoustic Reverberation Cancellation
Application of Bayesian methods to remove acoustic reverberation from audio signals, both in a single-channel and multi-channel environment.

Single Channel Signal Separation
Novel techniques for tackling the fundamental problem of separating multi-signals from a single data-record; still a largely unsolved problem.

Image Processing
Application of Bayesian methods to the restoration of photographic images, to compensate for degradations due to, for example, motion blur, out-of-focus blur, and atmospheric turbulence.

Ultrasound Image Restoration
Blind deconvolution in ultrasound imaging is an important problem which can improve the clarity of the image presented to the clinician.