Biomedical Engineering Reference
In-Depth Information
In MR, however, there is negligible risk associated with the acquisition, and
modern imaging provides techniques for acquiring large numbers of slices in
a short time. As a result of these differences, it is common for an MR image to
have a much larger field of view in the through-slice direction than the CT
image to which it is being registered.
Another important difference between MR and CT imaging is that,
although MR scanners can acquire images in arbitrary slice orientations, CT
scanners are limited to axial acquisitions or, by careful patient positioning
and maximum tilting of the gantry, to images of the head in the coronal plane.
It may, therefore, be desirable to align images acquired in different planes
with the two modalities; for example, to register sagittal MR images with
axial CT images. In this case, the overlapping portion of the image field of
view may be considerably smaller than either image. An example set of MR
and CT images of the head is shown in Figure 10.1.
The difference in field of view of MR and CT images and the restricted vol-
ume of overlap of the two images present considerable challenges to image
registration algorithms. In many early applications of MR-CT registration it
was essential, in order to get good registration accuracy, to acquire both images
using a similar field of view and as similar as possible a slice orientation.
FIGURE 10.1
Example MR and CT images of the head before registration. Note the difference in field of
view in the through-slice direction.
