A5088513060- Electrical and Bioimpedance Tomography
- Medical Imaging Techniques and Applications
- Fault Detection and Control Systems
- Geophysical and Geoelectrical Methods
- Advanced X-ray and CT Imaging
- Flow Measurement and Analysis
- Soil Geostatistics and Mapping
- Statistical Methods and Inference
- Statistical Distribution Estimation and Applications
- Advanced MRI Techniques and Applications
- Bayesian Methods and Mixture Models
- Non-Destructive Testing Techniques
- Image and Signal Denoising Methods
- Medical Image Segmentation Techniques
- Statistical and numerical algorithms
- Statistical Methods and Bayesian Inference
- Gaussian Processes and Bayesian Inference
- Forensic Anthropology and Bioarchaeology Studies
- Structural Health Monitoring Techniques
- Plant Water Relations and Carbon Dynamics
- Image and Object Detection Techniques
- Image Processing and 3D Reconstruction
- Advanced Statistical Process Monitoring
- Insurance, Mortality, Demography, Risk Management
- Geochemistry and Geologic Mapping
University of Leeds
2014-2025
Statistics Finland
2008-2013
Engineering and Physical Sciences Research Council
2001
University of Bradford
1994-2001
University of Bristol
1991
Accurate estimation of human adult age has always been a problem for anthropologists, archaeologists and forensic scientists. The main factor contributing to the difficulties is high variability physiological indicators. However, confounding this in many applications systematic tendency estimates, regardless indicator employed, assign ages which are too young individuals, low older individuals. This paper shows that at least part error inevitable consequence statistical procedures used...
Much of the data which appears in forensic and archaeological literature is ordinal or categorical. This particularly true age related indicators presented by Gustafson his method human adult estimation using structural changes teeth. technique still being modified elaborated. However, statistical methods regression analysis employed others are not appropriate to this type data, but because alternatives have yet been explored. paper presents a novel approach based upon application Bayes'...
It is generally assumed that life expectancy in antiquity was considerably shorter than it now. In the limited number of cases where skeletal or dental age-at-death estimates have been made on adults for whom there are other reliable indications age, appears to be a clear systematic trend towards overestimating age young adults, and underestimating older individuals. We show this might result use regression-based techniques analysis converting indicators into estimated ages. Whilst...
The use of high-power industrial equipment, such as large-scale mixing equipment or a hydrocyclone for separation particles in liquid suspension, demands careful monitoring to ensure correct operation. fundamental task state-estimation the suspension can be posed time-evolving inverse problem and solved with Bayesian statistical methods. In this article, we extend methods incorporate models error that is incurred numerical solution physical governing equations. This enables full uncertainty...
The recently introduced kernelized expectation maximization (KEM) method has shown promise across varied applications. These studies have demonstrated the benefits and drawbacks of technique when kernel matrix is estimated from separate anatomical information, for example magnetic resonance (MR), or a preliminary PET reconstruction. contribution this work to propose investigate list-mode-hybrid KEM (LM-HKEM) reconstruction algorithm with aim maintaining anatomically-guided methods overcome...
SUMMARY A method is proposed for the calibration of a continuous random variable when dependent variables are combination and categorical, model between controlling calibrated empirically derived. The various probability distributions estimated from training data by using kernel density procedures with bi-variate normal kernels uniform smoothing discrete variables. Bayes's theorem then used to produce posterior distribution which point estimates confidence may be made. Individual densities...
Abstract The Birnbaum–Saunders (BS) distribution is a model that frequently appears in the statistical literature and has proved to be very versatile efficient across wide range of applications. However, despite growing interest study this development many articles, few them have considered data with dependency structure. To fill gap, we introduce new class time series models based on BS distribution, which allows modeling positive asymmetric an autoregressive We call these moving average...
Medical images are hampered by noise and relatively low resolution, which create a bottleneck in obtaining accurate precise measurements of living organisms. Noise suppression resolution enhancement two examples inverse problems. The aim this study is to develop novel robust estimation approaches rooted fundamental statistical concepts that could be utilized solving several problems image processing potentially reconstruction. In study, we have implemented Bayesian methods been identified...
It is very common for multiple experiments to be conducted under non-identical but similar conditions, perhaps because of implementation errors, natural variability in experimental material, or gradual drifting conditions. In the extremes modelling, each dataset could analysed independently differences ignored entirely and datasets treated as replicates. this paper, an alternative approach proposed which a parametric family assumed across all datasets, then links parameters separate through...
After a brief review of the paradigm bayesian image restoration, we pose question: If high-level prior information is available and usable, what lost by modelling at pixel level instead? Our discussion based on real application where this question relevant: use gamma-camera imagery in location lesions. Procedures using ‘global' form structural model for are compared with those more conventional 'local' priors, only interactions among neighbouring pixels. We address detail physical...
Anatomically-driven image reconstruction algorithms have become very popular in positron emission tomography (PET) where they demonstrated improved resolution and quantification. This work, consider the effect of spatial inconsistency between MR PET images hot cold regions image. We investigate these effects on kernel method from machine learning, particular, hybrid kernelized expectation maximization (HKEM). These were applied to Jaszczak phantom patient data acquired with Biograph Siemens...
Understanding patterns in the frequency of extreme natural events, such as earthquakes, is important it helps prediction their future occurrence and hence provides better civil protection. Distributions describing these events are known to be heavy tailed positive skew making standard distributions unsuitable for modelling events. The Birnbaum–Saunders distribution its value version have been widely studied applied due attractive properties. We derive L-moment equations propose novel methods...
Abstract Over recent years, the use of homogeneous Gibbs prior models in image processing has become widely accepted. There been, however, much discussion over precisely which are most appropriate. For applications, simplest Gaussian model tends to oversmooth reconstructions, so it been rejected favor various edge-preserving alternatives. We claim that problem is not with family, but rather assumption homogeneity. In this article we propose an inhomogeneous random field as a general for many...
This paper investigates Bayesian estimation for Gaussian Markov random fields. In particular, a new class of compound model is proposed which describes the observed intensities using an inhomogeneous and degree spatial variation described by second field. The coupled fields are used as prior distributions, combined with noise models to produce posterior distributions on based. All parameters estimated, in fully setting, Metropolis-Hasting algorithm. full procedures illustrated compared...
Many imaging problems such as with electrical impedance tomography (EIT) can be shown to inverse problems: that is either there no unique solution or the does not depend continuously on data. As a consequence of based measured data alone unstable, particularly if mapping between distribution and measurements also nonlinear in EIT. To deliver practical stable solution, it necessary make considerable use prior information regularization techniques. The role Bayesian approach therefore...
Since its origins and numerous applications in material science, the Birnbaum–Saunders family of distributions has now found widespread uses some areas applied sciences such as agriculture, environment medicine, well quality control, among others. It is able to model varied data behaviour hence provides a flexible alternative most usual distributions. The includes log-Birnbaum–Saunders univariate multivariate versions. There are well-developed methods for estimation diagnostics that allow...