Activities in this research area are governed by Joost Batenburg

We develop algorithms and numerical solution techniques for large-scale inverse problems in imaging. Our approach is interdisciplinary, combining aspects of Mathematics (Numerical Linear Algebra, Discrete Mathematics,  Fourier Analysis, ...), Computer Science (Large scale parallel computing) and Physics (Image formation modeling). Our algorithms can be applied to a wide range of image reconstruction problems in science, medicine, and industry.

Current projects include:
3D imaging of nanocrystals at atomic resolution" (with S. Van Aert, S. Bals, and G. Van Tendeloo, Univ. of Antwerp)

We develop a workflow for 3D electron microscopy at atomic resolution, by combining advanced image acquisition, parameter estimation, and discrete tomography algorithms.

Ultra-fast iterative reconstruction" (with L. Plantagie)

Here, we develop algorithms for tomography that obtain similar reconstruction quality to time-consuming iterative methods, at a fraction of the computational cost.

Guaranteed bounds for reconstruction accuracy (with W.R. Fortes, R. Tijdeman, L. Hajdu)

We develop a mathematical framework for bounding the difference between an exact image of a scanned object and a computed reconstruction of the object.

Quantitative electron tomography by combined parameter estimation and reconstruction (NWO Vidi project, 2011-2016)

In this project we develop theory and algorithms for quantitative image reconstruction in electron tomography of nanomaterials. The research focuses on integrating the estimation of various acquisition and reconstruction parameters within the reconstruction algorithm.

Optimization of the structure-activity relation in porous nanomaterials (with P. Cool, S. Bals, B. Partoens, Univ. of Antwerp)

We develop an extensive methodology for optimized design of nanomaterials that combines novel techniques for material synthesis with advanced image acquisition, reconstruction and simulation methods.

Dynamic angle selection for tomography (with P. Balazs, Univ. Szeged, Hungary)

Here, we develop theory and algorithms for dynamic selection of scanning angles in tomography, that maximize the expected information gain.

Super-CT: Speeded Up Processing and Reconstruction of Tomograms (with J. Sijbers, Univ. of Antwerp and other partners)
This project aims at the development of high-performance parallel implementations of tomographic reconstruction algorithms, with an emphasis on GPU computing.