Dr Robert G. Aykroyd
My research interests are in the general area of applied statistics. Perhaps the main area being the use of Bayesian modelling approaches in image analysis and other inverse problems. Usually this involves the use of hyper-priors and prior parameter estimation. This has involved applications in archaeological geophysics, medical imaging (including PET & SPECT) and engineering applications of electrical tomography. Often estimation involves use of MCMC algorithms. I also have a strong interest in use of other numerical techniques such as finite element and boundary element methods. More fundamental work has looked at inhomogeneous Markov random field models, 'phase change' properties of spatial auto-models, and approximations of normalizing constants in certain Markov random field models. I also have interest in Wavelet methods and kernel density methods. These have been applied in process monitoring, image analysis, forensic age estimation and climate reconstruction. I would like to start new work on variational Bayesian methods for inverse problems. If you are interested, then please get in touch.
Research groups and institutes
<li><a href="//phd.leeds.ac.uk/project/639-bayesian-modelling-for-medical-image-reconstruction">Bayesian modelling for medical image reconstruction</a></li>
<li><a href="//phd.leeds.ac.uk/project/245-fast-bayesian-estimation-using-expectation-propagation-methods-applied-to-inverse-problems">Fast Bayesian estimation using expectation propagation methods applied to inverse problems</a></li>
<li><a href="//phd.leeds.ac.uk/project/204-locally-adaptive-bayesian-modelling-for-medical-image-reconstruction">Locally-adaptive Bayesian modelling for medical image reconstruction</a></li>