title: Artifact Correction and Real-Time Scatter Estimation for X-Ray Computed Tomography in Industrial Metrology
creator: Maier, Joscha
description: Artifacts often limit the application of computed tomography (CT) in industrial metrology. In  order to correct these artifacts, the so-called simulation-based artifact correction (SBAC) was  developed in this thesis. For this purpose, analytical and Monte Carlo (MC) based models  were set up to simulate the CT measurement process for a given component as accurately  and efficiently as possible. Calculating the difference between this simulation and an ideal  one yields an estimate of the present artifacts that can be used to correct the corresponding  CT measurement. The potential of this approach was demonstrated for the correction of the  most common CT artifacts, i.e. beam hardening, x-ray scattering, off-focal radiation, partial  volume effects, and cone-beam artifacts. In any case, the SBAC provided CT reconstructions  that showed almost no artifacts and whose quality was clearly superior to state-of-the-art  reference approaches. In this context, the problem of long runtimes of scatter simulations was  solved by another novel approach, the so-called deep scatter estimation (DSE). The DSE uses  a deep convolutional neural network which was trained to map the acquired projection data  to given MC scatter estimates. Once the DSE network is trained, it can be used to process  unknown data in real-time. In different simulation studies and measurements, it could be  shown that DSE generalizes to various acquisition conditions and components while providing  scatter distributions that differ by less than 2 % from MC simulations. Thus, the two developed  approaches make an important contribution to correct CT artifacts efficiently and to extend the  applicability of CT in the field of industrial metrology.
date: 2019
type: Dissertation
type: info:eu-repo/semantics/doctoralThesis
type: NonPeerReviewed
format: application/pdf
identifier: https://archiv.ub.uni-heidelberg.de/volltextserverhttps://archiv.ub.uni-heidelberg.de/volltextserver/26701/1/Dissertation_JoschaMaier.pdf
identifier: DOI:10.11588/heidok.00026701
identifier: urn:nbn:de:bsz:16-heidok-267010
identifier:   Maier, Joscha  (2019) Artifact Correction and Real-Time Scatter Estimation for X-Ray Computed Tomography in Industrial Metrology.  [Dissertation]     
relation: https://archiv.ub.uni-heidelberg.de/volltextserver/26701/
rights: info:eu-repo/semantics/openAccess
rights: http://archiv.ub.uni-heidelberg.de/volltextserver/help/license_urhg.html
language: eng