Biomedical Engineering Reference
In-Depth Information
8.9 Mass conservation{based optical ow
The above mentioned methods for motion correction are applicable to the
PET data only when correspondences in the gray values of the target and
floating images are given. In cardiac motion, this is not the case. The so-
called partial volume effect (PVE) becomes prominent in cardiac studies. It is
a result of the limited resolution of the scanners. All objects smaller than the
scanner resolution limit cannot be accurately delimited and therefore appear
blurred. As the heart muscle contracts and expands during the cardiac cycle
its thickness varies. In phases with thicker heart wall, the activity is better
resolved and has a higher amplitude as compared to other phases where the
myocardium is thin (see Figure 8.6). However, the total amount of the activity
remains the same.
Two important studies related to the correction for cardiac motion are [36]
and [31]. In the first study optical flow is used for estimating the deforma-
tions in the images by modeling the myocardium as an elastic membrane. The
second study combines the motion estimation of the first study with recon-
struction in a single framework. However this study is confined to 2D images
and deals with cardiac motion.
The method presented below is essentially different from the brightness
FIGURE 8.6: Two phases from the cardiac cycle of the heart. Above: end-
systole, Below: end-diastole. A coronal slice from the 3D PET image volume
is shown. Images from an FDG study are shown here without attenuation
correction.
Search WWH ::
Custom Search