Medical Imaging

Particularly positron emission tomography (PET) and single photon emission computerised tomography (SPECT). The main interests are
(1). Image reconstruction algorithms for maximum likelihood (ML) or maximum penalised likelihood (MPL) image reconstruction.
(2). Block iterative reconstruction methods.

Statistical Inversion Methods

Interested in solving those inverse problems where the noise distribution may not be follow normal distributions. Particularly interested in the algorithms for inverse problems with general exponential family noise distributions.

Missing Values and EM related algorithms

Expectation maximization (EM) method is a general approach for maximum likelihood parameter estimation, especially when there are missing values. However in many applications the Expectation step (E-step) is intractable. An algorithm, which also uses augmented data, is developed. It does not require the conditional expectation computation for E-step and its convergence behaviour is very similar to EM.

Other Interests

Other research interest includes: non-parametric smoothing, random effects models, survival analysis, censored regression.