Bayesian Estimation of Density via Multiple Sequential Inversions of Two-Dimensional Images With Application to Electron Microscopy

Dalia Chakrabarty, Nare Gabrielyan, Fabio Rigat, Richard Beanland, Shashi Paul
  • Technometrics, April 2015, Taylor & Francis
  • DOI: 10.1080/00401706.2014.923789

A harder-than-usual inverse problem

What is it about?

The novel methodology presented herre using image synthesis ideas and newly developed priors on sparsity, within a very high-dimensional MCMC-based inferential scheme.

Why is it important?

We present a new Bayesian methodology to learn the unknown material density of a given sample by inverting its two-dimensional images that are taken with a Scanning Electron Microscope. The method is applied to learn the material density of a 3-D sample of a nano-structure, using real image data. Illustrations on simulated image data of alloy samples are also included.

Perspectives

Professor Shashi Paul
De Montfort University

Working on this article with mathematicians has given me different perspective of solving a problem using both mathematical and computation tools. Designing the experiment was a real challenge.

Read Publication

http://dx.doi.org/10.1080/00401706.2014.923789

The following have contributed to this page: Dr Dalia Chakrabarty and Professor Shashi Paul