UNIVERSITY OF ST ANDREWS
SCHOOL OF MATHEMATICS AND STATISTICS
STATISTICS SEMINAR PROGRAMME: AUTUMN 2002
Statistics seminars will be held at 4.00 pm on Mondays in
Lecture Theatre D on the top floor of the Mathematical Institute. Tea
and biscuits are available in the Staff Room 20 minutes before the
seminars.
14 October Dr. Elke Thönnes (University of Warwick)
"Inferring global vein/artery structure in medical
angiograms using Markov
chain Monte Carlo"
11 November Dr. Chris Glasbey (Biomathematics and Statistics
Scotland, Edinburgh)
"Analysis of microarrays and electrophoresis gels"
2 December Dr. Nicole Augustin (University of Glasgow)
"Spatial modelling of different types of forestry data"
============================================================
ABSTRACTS:
Inferring global vein/artery structure in medical angiograms using
Markov chain Monte Carlo
(Dr. Elke Thönnes)
Detecting vein and artery structure in medical angiograms is an
important problem for surgical planning. In this talk I describe a
method for inferring tree-like, vascular structure in 2D imagery.
First, a multi-resolution approach is used to determine local
vascular structure. Based on the resulting image summarisation and a
prior distribution that models random trees, a posterior distribution
for the global vascular structure is formulated and sampled using an
MCMC algorithm. Thus, combining global statistical modelling with
efficient local feature extraction maintains statistical
interpretability whilst ensuring efficiency of the resulting
algorithm. The method is illustrated on examples of retinal
angiograms.
-------------------------------------------------------------------
Analysis of microarrays and electrophoresis gels
(Dr. Chris Glasbey)
Micorarrays and gel electrophoresis are key technologies in genomics,
and beyond. Statistical models will be presented for automating the
identification and quantification of spots or bands for a range of
problems. Methods take account of noise, spatial warping, image
segmentation and use of reference datasets, and will be illustrated.
-------------------------------------------------------------------
Spatial modelling of different types of forestry data
(Dr. Nicole Augustin)
In this talk two data-analysis examples from forestry are presented,
where the focus is on estimating effects of explanatory variables in
spatial data, rather than prediction. We account for the spatial
correlation in the data using different approaches including
generalised estimation equations, generalised additive models and
generalised linear/additive mixed models.
In the first example the spread of Nectria canker of beech, a fungal
chronic disease, is investigated. Data are available from a beech
provenance trial. A possible influential factor is the wind
dispersal zone, a categorical variable describing the distance and
wind direction from diseased shelterwood, the source of infection.
We estimate its effect on the proportion of infected trees per plot.
In the second example it is the aim to identify a set of suitable
explanatory variables (from a large set) for modelling defoliation of
pine trees. The data are from a monitoring survey carried out in 1994
in the forest of Baden-Württemberg, Germany.
==================================================================
Please check http://www-maths.mcs.st-and.ac.uk/StatsSeminars/ shortly
before coming to a seminar, in case there are last-minute changes,
cancellations, etc.
******************************************************************
Dr. P. E. Jupp
School of Mathematics and Statistics
University of St. Andrews
North Haugh, St. Andrews tel: (44) 1334 463704
Fife, KY16 9SS fax: (44) 1334 463748
Scotland e-mail: [log in to unmask]
******************************************************************
|