Affiliations: Biomedical Physics, Department of Physics, University
of Antwerp, 171 Groenenborgerlaan, 2020 Antwerp, Belgium
Note: [] Corresponding author: Stefan L.R. Gea, Laboratory of Biomedical
Physics (BIMEF), Department of Physics, University of Antwerp,
Groenenborgerlaan 171, B-2020 Antwerpen, Belgium. Tel.: +32 3 265 34 38; Fax:
+32 3 265 33 18; E-mail: stefan.gea@ua.ac.be
Abstract: Computed microtomography (μCT) is an imaging
technique that produces virtual slices of an object, based on a set of x-ray
images taken while the object is stepwise rotated. The stack of slices is
calculated with a backprojection algorithm, it demands that the entire object
remains within the x-ray beam for all observed angles. During
region-of-interest (ROI) tomography however, one zooms in on a small part of
the object. As a consequence the above condition is no longer fulfilled; the
input data become incomplete and reconstructed slices are corrupted. In this
study we show that, if we are not interested in grayscale values but only in
geometrical information, conventional backprojection algorithms still perform
well for ROI tomography. We demonstrate this with ROI scans performed on
several phantom objects and we show ROI images of the gerbil hearing organ. We
found that the resolution could considerably be improved and small details, not
visible in a conventional full object scan, can be revealed by a ROI scan. One
can use the benefits of ROI-μCT in situations where
physical constrains are such that large parts of the object exceed the field of
view (FOV), or where a very high magnification is required.
Keywords: Microtomography, region of interest, temporal bone, middle ear
