Biomedical Engineering Reference
In-Depth Information
the circular region are compared with the adjacent image (next projection) using the
sum of squared differences, defined by:
X
X
I x k I y ` I 0 i k I j ` 2
SS D ij D
;
(20)
k
`
.x; y/ are the coordinates of the circular region center, .k; `/ are in the range of
pixels in the region of interest, I is the original image and I 0 the adjacent image.
The minimum of the function SSD ij is calculated by a parabolic interpolating and
the coordinates of the minimum and .x; y/ are the translation values that align
the image I 0 with the image I . The center of the circular region is then placed in the
adjacent image, I 0 at the coordinates .x C x; y C y/ and the process repeats for
a new adjacent image. At the end all images will be aligned.
ME techniques are perhaps the most straightforward since motion is computed
from the deviations of the external markers centroid seen in different projection
images.
Motion in the OF algorithm is obtained by optical flow (movement of an
intensity pattern on the image caused by the movement of the scene or the camera)
considering two consecutive projections. For a faster calculation optical flow is
generally determined on a region of interest. This option permits also to reduce the
possible effects of motion on the determination of optical flow since it is assumed
that these effects are global.
In addition to the methods mentioned before, the visualization of the projection
images consecutively as if it were a film enable to detect sudden movements.
Several studies comparing [ 51 - 55 ] the various methods were carried out but there
are no agreement in the conclusions and it seems that there is not a method that
performed better than all the others.
3.2.2
Motion Detection Using External Devices
Motion tracking is technically possible with various types of transducers, however
the requirements demanded at the clinical setting significantly limits the available
solutions. Therefore, video cameras working either in the visible [ 56 - 58 ]orinthe
near-infrared range [ 59 - 61 ] have been the most frequent option. The basic idea of
these methods is to place markers on the patient's body that could be seen by various
cameras and that could be tracked three-dimensionally. Following the marker's
positions it is possible to measure the motion of the body which they are linked
to. Cameras working in near-infrared range allow to operate with dark conditions
that are advantageous for some neurological examinations, and also allow better
markers identification since the signal to noise ratio is higher. On the other hand, the
use of cameras working in the visible range allows a reduction of costs and the use
of a large number of markers.
Search WWH ::




Custom Search