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.