On the precision of rotation axis estimation in Computed Tomography

presentation

English

Misalignment of the axis of rotation (AoR) in computed tomography (CT) is known to cause characteristic tuning fork artifacts or double edges in CT slices. Various automatic methods are commonly used to estimate and correct the AoR position before reconstruction. These methods aim for high accuracy and precision, but sources differ on the required level of precision, i.e. the amount of AoR shift where errors become noticeable. Some works estimate the magnitude of AoR misalignment artifacts analytically, but a practical assessment of how different factors in a typical CT scan influence the precision is not readily available in the English literature. Such an assessment can be useful when tuning or assessing an AoR estimation methods, as different scan scenarios can require different levels of precision and therefore different amounts of time and effort needed for tuning the AoR. Here, we experimentally assess how common factors such as noise, sample stage instability, sample contrast, and geometry of the CT system influence the required precision of AoR estimation and correction. This will hopefully shed light on the well-known, but somewhat understudied AoR misalignment artifacts.

computed tomography;axis of rotation;artifacts

ZEMEK, M.; ŠALPLACHTA, J.; ZIKMUND, T.; OBERTA, P.; KAISER, J.

29. 6. 2022

Grenoble

1

1

1

https://ictms2022.sciencesconf.org/