Edinburgh Research Explorer

Michael Davies

Jeffrey Collins Chair of Signal Processing

Profile photo

Willingness to take PhD students: Yes

Compressed sensing and sparse representations in signal processing and machine learning

Education/Academic qualification

Doctor of Philosophy (PhD), University College London
Noise Reduction in Nonlinear Time Series Analysis
Bachelor of Arts, University of Cambridge

Professional Qualifications

2000Chartered Engineer, CIET


Career History

Mike Davies holds the Jeffrey Collins Chair in Signal and Image Processing at the University of Edinburgh where he leads the Edinburgh Compressed Sensing Research Group. He was Head of the Institute for Digital Communications (IDCom) in the School of Engineering 2013-16 and Director of the Joint Research Institute in Signal and Image Processing (JRI-SIP) a the collaborative research venture between the University of Edinburgh and Heriot-Watt University as part of the Edinburgh Research Partnership.

He received an M.A. in engineering from Cambridge University in 1989 where he was awarded a Foundation Scholarship (1987), and a Ph.D. degree in nonlinear dynamics and signal processing from University College London (UCL) in 1993. In 1993 he was awarded a Royal Society University Research Fellowship. He currently manages a £7M portfolio of research grants from a variety of sources including: EPSRC, Dstl, industry, EU and the ERC, and leads the University Defence Research Collaboration (UDRC), a UK programme of signal processing research in defence in collaboration with the UK Defence Science and Technology Laboratory (Dstl).

Awards and Fellowships

Fellow of the Royal Society of Edinburgh (2018)

Fellow of the Royal Academy of Engineering (2017)

Royal Society Wolfson Research Merit Award (2016)

Fellow of the European Association of Signal Processing (2016)

Fellow of the IEEE (2015)

Texas Instruments Distinguished Visiting Professor, Rice University (2012)


Research Interests

My work focusses on the fields of sparse representations and compressed sensing, and their application to various signal processing, imaging and machine learning problems. During the past decade I have made significant contributions to the key areas of fundamental CS theory, new signal models, including dictionary learning techniques, the development and analysis of new reconstruction algorithms, This work includes: the proposal and analysis of the highly popular Iterative Hard Thresholding family of algorithms for sparse reconstruction; the development of new reconstruction theory for structured sparse signal models; the introduction and analysis for a new model (co-sparsity) for redundant analysis representations; and the characterization of types of statistical distribution that admit accurate low dimensional approximations.

I have applied these ideas to a number of applications including: chemical identification in Raman spectroscopy, dynamic MRI and 3D brain imaging, a new compressed sensing framework for quantitative MRI, electronic surveillance and Synthetic Aperture Radar.

Current Projects

ERC Advanced Grant: C-SENSE, "Exploiting Low Dimensional Models in Sensing, Computation and Processing"

The aim of this project is to develop the next generation of compressive and computational sensing and processing techniques.

UDRC: University Defence Research Collaboration in Signal Processing

An academia led partnership between the defence industry, academia and the government sector. The UDRC develops research in signal processing with application to the defence industry.

Recent Past Projects

CQ-MRI: EPSRC funded award in Compressed Quantitative MRI

The proposed research will provide the first proof-of-principle for a new family of Compressed Quantitative Magnetic Resonance Imaging (CQ-MRI), able to rapidly acquire a multitude of physical parameter maps for the imaged tissue from a single scan.

CIRI: EPSRC funded award in Compressed Imaging in Radio Interferometry

The CIRI project aims to bring new advances for interferometric imaging for next-generation radio telescopes, together with theoretical and algorithmic evolutions in generic compressive imaging.

EU Innovative Training Networks:

SpaRTaN: Sparse Representations and Compressed Sensing Training Network

MacSeNet: Machine Sensing Training Network

Professional Activities

2013 – date         Member of IEEE Sig. Proc. Theory and Methods technical committee


ID: 9466