Number of classes
|
The number of tissue classes to estimate. The default value is three,
representing the Cerebrospinal fluid, Grey matter, and White Matter.
|
Correct Inhomogeneity
|
With this enabled FANTASM will try to estimate the inhomogeneity field
within the optimization process.
|
Background Cropping
|
Crops background values from being included in the estimation problem.
|
Output Type
|
Sets the desired outputs.
| all_result_images | Outputs all image types
(fuzzy segmentation, hard segmentation, inhomogeneity field). |
| hard_segmentation | Outputs only the hard
segmentation based on thresholding the fuzzy segmentation. |
| fuzzy_segmentation | Outputs only the
fuzzy segmentation. |
| both_fuzzy_&_hard | Outputs both the hard
segmentation and the fuzzy segmentation. |
|
Smoothing parameter
|
This determines the smoothness of the fuzzy segmentation. Higher
values will make the segmentation smoother, while values of zero mean
no smoothing.
|
Maximum difference
|
This controls the maximum difference between successive iterations of
the algorithm which can cause the algorithm to terminate. In essence
if the change in any of the FCM centroids is bigger than this number,
then the algorithm will continue to iterate until the change is below
this number or the maximum number of iterations happens.
|
Maximum iterations
|
Specifies the maximum number of iterations allowed before the process
is terminated. The default value is 50.
|
Initialization
|
How to initialize the algorithm.
| range | Takes the image range and divides it
into the same number of intervals as classes, and sets the initial
centroids to be the center value of these intervals. |
| modes | Selects modes based on subdividing the
histogram to find the N most common modes, where N is the desired
number of classes. |
| manual | Allows users to manually enter
initial centroid values. |
|
Masking Mode
|
|
Inhomogeneity field degree
|
If Correct Inhomogeneity is on, this value determines the degree of the
B-Spline used to estimate the inhomogeneity field.
|
Background threshold
|
If the Background Cropping is on, this values determine what
the algorithm will regard as background. That is, all values between
this value and the minimum intensity within the image will be thought
of as background.
|
Fuzziness Coefficient
|
Controls the fuzziness of the FCM model. Smaller values are less
fuzzy, a value of 1.0 represents no fuzziness.
|
