NAME
Atropos - part of ANTS registration suite
DESCRIPTION
COMMAND:
./Atropos
Atropos:
A priori classification with registration initialized template
assistance.
OPTIONS:
-i, --initialization:
Option 1:
Random[numberOfClasses]
Option 2:
Kmeans[numberOfClasses]
Option 3:
Otsu[numberOfClasses]
Option 4:
PriorProbabilityImages[numberOfClasses,fileSeriesFormat(index=1
to numberOfClasses) or vectorImage,priorWeighting]
Option 5:
PriorLabelImage[numberOfClasses,labelImage,priorWeighting]
-a, --intensity-image:
[intensityImage,<adaptiveSmoothingWeight>] -- adaptive smoothing
only applies to initialization with prior image(s)
-x, --mask-image:
maskImage
-c, --convergence:
[<numberOfIterations>,<convergenceThreshold>]
-k, --likelihood-model:
Option 1: Gaussian Option 2:
ManifoldParzenWindows[<pointSetSigma=1.0>,<evaluationKNeighborhood=50>,<CovarianceKNeighborhood=0>,<kernelSigma=0>]
Option 3: HistogramParzenWindows[<Sigma=1.0>,<numberOfBins=32>]
-m, --mrf:
[<smoothingFactor>,<radius>]
-o, --output:
[classifiedImage,<posteriorProbabilityImageFileNameFormat>]
-u, --minimize-memory-usage:
minimize-memory-usage=1/(0) <VALUES>: 0
-b, --bspline:
[<numberOfLevels>,<initialMeshResolution>,<splineOrder>] -- only
applies to initialization with prior image(s) and non-zero
adaptive smoothing parameter.
-e, --use-euclidean-distance:
use euclidean or geodesic distance for prior label distance maps
= 1/(0) -- only applies to initialization with a prior label
image. <VALUES>: 0
-w, --winsorize-outliers:
Option 1:
BoxPlot[<lowerPercentile=0.25>,<upperPercentile=0.75>,<whiskerLength=1.5>]
Option 2:
GrubbsRosner[<significanceLevel=0.05>,<winsorizingLevel=0.10>]
-l, --labels:
whichLabel[sigma,<boundaryProbability>] -- only applies to
initialization with a prior label image.
-h, --help:
Print menu. <VALUES>: 1, 0