My research programme is focused on Bayesian inference and scalable computation for emerging applications in multi-spectral and hyper-spectral imaging. As the spectral and spatial resolutions of instruments have improved, there has been an increasing need in analytical chemistry to process large volumes of complex data. Raman mapping in 2D and 3D enables nanometrology and imaging of biological processes at the molecular level. I am working on model-based approaches for source separation and quantification in this context. My sequential Monte Carlo (SMC) algorithm is available in the R package serrsBayes. In previous work, I have developed accelerated algorithms for approximate Bayesian computation (ABC) and pseudo-marginal methods using surrogate models. A recent preprint is available here. I have implemented these algorithms in my R package bayesImageS, available on CRAN. I give a brief introduction to this research in this YouTube video, where I demonstrate image segmentation for satellite remote sensing and cone-beam computed tomography (CT).
Härkönen, T., Roininen L., Moores M., & Vartiainen E. M.
(2020). Bayesian Quantification for Coherent Anti-Stokes Raman Scattering Spectroscopy. The Journal of Physical Chemistry B. doi: 10.1021/acs.jpcb.0c04378
Hargrave, C., Mason N., Guidi R., Miller J-A., Becker J., Moores M., et al.
(2016). Automated replication of cone beam CT-guided treatments in the Pinnacle treatment planning system for adaptive radiotherapy. Journal of Medical Radiation Sciences. 63(1), 48-58. doi: 10.1002/jmrs.141
Falk, M. G., Alston C. L., McGrory C. A., Clifford S., Heron E. A., Leonte D., et al.
(2015). Recent Bayesian approaches for spatial analysis of 2-D images with application to environmental modelling. Environmental and Ecological Statistics. 22(3), 571-600. doi: 10.1007/s10651-015-0311-1