|
| |
Restoration Dialog
  
Use this dialog to perform
restoration on images to increase the Signal to Noise Ratio.
Note
Most of the settings for the Restoration Dialog can be set through
a resource file
that is loaded at the start of the program.
 |
Example performing restoration on
an image set (reference) |
-
Load all sequences from a
single location from a patient study that are in alignment. If the
images are not aligned, run Registration.
-
Select
Automatic in the Region of Interest/Volume of
Interest Dialog.
-
Press the left button over the white matter area contralateral to the abnormal anatomy.
-
Select
Current Image from Region or Volume
Operation and press Stats.
Note
The statistics should be for an individual image, do not use the All button.
-
Calculate 2 x Sigma Avg.
which will appear in the Information Window, and enter
this value for
sigma in the Restoration Window.
-
Press
Clear in the Region of
Interest/Volume of Interest Dialog.
-
Press
Restore to perform restoration on the images.
-
Check the resulting images for blurring. If images are
blurry, reduce sigma and repeat Step
7. This can be repeated until
the resulting images are satisfactory.
-
Once the
restoration is successful, press
Repeat in the Restoration Dialog. Select a range of locations, if
desired, ensuring this range contains usable images, and Press OK in
the Repeat Dialog.
|
Dialog Items
Note
The following items may or may not appear, depending on your
selection of Restoration Method.
 |
# Images
Select n,
the # Images, to use from the images loaded into the selected
browser. Only the last n images will be used to perform the
restoration.
|
|
# Out
Select n
from # Out, the number of images to output. Interpolation is
used to create the extra images in between.
|
|
Extra Output
 |
Select None
to create only the restored images.
|
|
Select Cor Coef
to create a correlation coefficient image along with the restored
images.
|
|
Select # Avg
to create an image of the number of pixels averaged to create each pixel,
along with the restored images.
|
|
|
Restore
Press Restore
to perform the restoration, on each pixel vector, of the images in the
selected browser.
|
|
Restoration
Method
|
Select MLE
to perform restoration using a maximum likelihood estimate to perform the
restoration.
Warning
Do not use this method when using restoring less than 3 images.
|
|
Select Kalman
to perform restoration using a Kalman filter.
|
Enter the P
value in P(0).
|
|
Enter the R
value in R(i) for each image.
|
Press Sel R
to calculate the R value on each image using the current region of interest.
|
|
|
Press Set Alpha
to set the alpha values for the pixels in the current region of interest.
|
Enter a value
for Alpha.
|
|
Note
This restoration method has not yet been implemented.
|
|
Select Average
to perform restoration using the neighborhood average.
|
|
Select Median to
perform restoration using the neighborhood median.
|
|
Select Average, Median
to use the best fit between Average or Median on a per pixel basis.
|
|
Select Average, MLE
to use the Average method on all the images, followed by the MLE method
on the first four.
Note
This restoration method has not yet been implemented.
|
|
Select Butterworth
to perform restoration using a butterworth filter with cosine extrapolation.
|
Enter a value
of the Butterworth filter in Order.
|
|
Enter the
cutoff frequency for the Butterworth filter in Cutoff.
|
|
|
Select Exponential
to perform restoration using the best fit to an exponential curve.
|
|
Select 2-Gaussian
to perform restoration using the best fit to 2 gaussian curves.
|
|
Select Gaussian-Exponential
to perform restoration using the best fit to a gaussian and exponential
curve.
|
|
Select LMRQ
to perform restoration using the Levenberg-Marquardt method.
|
Gamma
|
|
Gaussian
|
|
|
Select Spline
to perform restoration using a spline fit.
|
|
Select F(t)
to perform equation analysis on the selected images. Images will be created
that fit the original curve to a model of the passage of the bolus through
the brain.
|
|
Select -ln (S(t)/So)
to calculate the -ln (S(t)/So) on the images in the selected browser.
|
|
|
Skip
Select n,
from Skip to ignore pixels within n pixels of the center pixel.
The center pixel is never ignored.
|
|
X Function
|
Select Linear
to use linear value for X in the restoration. 1, 2, 3, .., n where n
is the number of images will be used for the X values.
|
|
Select Non-Linear
to use non-linear values for X in the restoration. The user supplies
the values for X in X(i).
|
|
Select Diffusion
to use the gradient (squared) for the X values in the restoration. The
user supplies values for S in S(i) to be converted to
gradient values.
|
|
Select T2
to use TE values for X in the restoration. A constant TR is assumed.
The user supplies the values for TE in TE(i) and TR in
TR, if the values cannot be determined from the image
header information. The output will be the T2 slope.
|
|
Select T1-Original
to use TR values for X in the restoration using the original method. A
constant TE is assumed. The user supplies the values for TR in
TR(i) and TE in TE, if the
values cannot be determined from the image header information. The
output will be the T1 slope.
|
|
Select T1-Eigen
to use TR values for X in the restoration using the eigen method. A
constant TE is assumed. The user supplies the values for TR in
TR(i) and TE in TE, if the
values cannot be determined from the image header information. The
output will be the T1 slope.
|
|
|
Y Function
|
Select Linear
to use the pixel values for the Y values.
|
|
Select Log
to use the log of the pixels values for the Y values.
|
|
|
TR, TE, X, S, B
|
Enter recovery
time in TR(i) or TR. TR(i) is used to specify a
recovery time for each image. TR is used to specify a constant
recovery time.
|
|
Enter echo time
in TE(i) or TE. TE(i) is used to specify an echo time
for each image. TE is to specify a constant echo time.
|
|
Enter a value
in X(i) for each image.
|
|
Enter the S
value in S(i) for each image.
|
|
Enter the B
value in B(i) for each image.
|
Note
These will only be displayed if needed and cannot be determined by the image
header information. See X Function for
details on which values may be needed.
|
|
Neighborhood
The neighborhood
is defined as the pixels around a central pixel that satisfy the following
constraints. If a pixel is in the neighborhood, it is used to
calculate the average pixel vector for calculating the restoration.
Also, that pixel's restoration results can be averaged with the center
pixel's restoration results. See
Estimation Method for how the results are averaged over the
neighborhood.
|
Neighborhood
size
|
Select the size,
in pixels, of the neighborhood in the X direction, centered on
a pixel.
|
|
Select the size,
in pixels, of the neighborhood in the Y direction, centered on a
pixel.
|
|
Select the size,
in pixels, of the neighborhood in the Z direction, centered on a
pixel.
|
|
|
Radius
Select Radius
to use an elliptical/spherical shape for the neighborhood, instead of a
rectangular/cubic shape.
|
|
sigma
Enter the
value for the euclidian distance between the center and non-center pixel
in sigma. If the euclidian-distance for a non-center pixel is
less than sigma, that pixel is part of the neighborhood.
Note
The default is no limit.
|
|
|
Estimation
Method
|
Select First Estimate
to use the restoration results from the first pixel, within the
neighborhood.
|
|
Select Center
Estimate to use the restoration results from the center pixel.
|
|
Select Mean of All
Estimates to average the restoration results from all pixels, within the
neighborhood.
|
|
|
Weight
|
Select None
to use no weighting in the fit of a pixel vector.
|
|
Select Std Dev
to use the standard deviation of the current region of interest, on each
image, as the weights in the fit of the pixel vector.
|
|
Select Pixel
to use the pixel intensity of each image as the weights in the fit of the
pixel vector.
|
|
|
Repeat
Press Repeat
to perform restoration on other browsers or files.
Note
Previous results of repeat processing need to be deleted
manually for images saved in Dicom format. |
|
Cancel
Press Cancel
to remove the Restoration dialog. |
|
Help
Press Help
to display help for the Restoration dialog. |
|
Back to
Eigentool
|
