brainscdetector:Usage

From NITRC Wiki

Return to the BRAINS Constellation Detector main page.

In this page, we will show the usage of the three modules in our BRAINSConstellationDetector project. They are the modeler, the detector, and the GUI corrector.

Modeler

BRAINSConstellationModeler 
[--returnparameterfile
<std::string>]
[--processinformationaddress
<std::string>] [--xml] [--echo] [-w
<int>] [-z <std::string>] [-y
<std::vector<int>>] [-x <double>]
[-n] [-q <int>] [-r <std::string>]
[-j <std::string>] [-l] [-m
<std::string>] [-i <std::string>]
[-v] [--] [--version] [-h]

Where:

 --returnparameterfile <std::string>
Filename in which to write simple return parameters (int, float,
int-vector, etc.) as opposed to bulk return parameters (image,
geometry, transform, measurement, table).

 --processinformationaddress <std::string>
Address of a structure to store process information (progress, abort,
etc.). (default: 0)

 --xml
Produce xml description of command line arguments (default: 0)

 --echo
Echo the command line arguments (default: 0)

 -w <int>,  --writedebuggingImagesLevel <int>
This flag controls if debugging images are produced.  By default value
of 0 is no images.  Anything greater than zero will be increasing
level of debugging images. (default: 0)

 -z <std::string>,  --BackgroundFillValue <std::string>
Fill the background of image with specified short int value. Enter
number or use BIGNEG for a large negative number. (default: 0)

 -y <std::vector<int>>,  --rescaleIntensitiesOutputRange
<std::vector<int>>
This pair of integers gives the lower and upper bounds on the signal
portion of the output image.  Out-of-field voxels are taken from
BackgroundFillValue. (default: 40,4000)

 -x <double>,  --trimRescaledIntensities <double>
Turn on clipping the rescaled image one-tailed on input.  Units of
standard deviations above the mean.  Very large values are very
permissive.  Non-positive value turns clipping off.  Defaults to
removing 0.00001 of a normal tail above the mean. (default: 4.4172)

 -n,  --rescaleIntensities
Flag to turn on rescaling image intensities on input. (default: 0)

 -q <int>,  --mspQualityLevel <int>
Flag cotrols how agressive the MSP is estimated.  0=quick estimate (9
seconds), 1=normal estimate (11 seconds), 2=great estimate (22
seconds), 3=best estimate (58 seconds). (default: 2)

 -r <std::string>,  --resultsDir <std::string>
The directory for the results to be written. (default: ./)

 -j <std::string>,  --optimizedLandmarksFilenameExtender <std::string>
If the trainingList is (indexFullPathName) and contains landmark data
filenames [path]/[filename].fcsv ,  make the optimized landmarks
filenames out of [path]/[filename](thisExtender) and the optimized
version of the input trainingList out of
(indexFullPathName)(thisExtender) , when you rewrite all the landmarks
according to the saveOptimizedLandmarks flag. (default: _opt.fcsv)

 -l,  --saveOptimizedLandmarks
Flag to make a new subject-specific landmark definition file in the
same format produced by Slicer3 with the optimized landmark (the
detected RP, AC, and PC) in it.  Useful to tighten the variances in
the ConstellationModeler. (default: 0)

 -m <std::string>,  --outputModel <std::string>
The full filename of the output model file. (default: default.mdl)

 -i <std::string>,  --inputTrainingList <std::string>
Setup file, giving all parameters for training up a template model for
each landmark.

 -v,  --verbose
Show more verbose output (default: 0)

 --,  --ignore_rest
Ignores the rest of the labeled arguments following this flag.

 --version
Displays version information and exits.

 -h,  --help
Displays usage information and exits.

Detector

 BRAINSConstellationDetector
[--returnparameterfile
<std::string>]
[--processinformationaddress
<std::string>] [--xml] [--echo] [-w
<int>] [-r <std::string>]
[--writeBranded2DImage
<std::string>] [-v] [--debug]
[--rVN4 <double>] [--rpc <double>]
[--rac <double>] [--rmpj <double>]
[--forceHoughEyeDetectorReportFailu
re] [--inputLandmarksEMSP
<std::string>] [--forceRPPoint
<std::vector<float>>]
[--forceVN4Point
<std::vector<float>>]
[--forcePCPoint
<std::vector<float>>]
[--forceACPoint
<std::vector<float>>]
[--interpolationMode
<NearestNeighbor|Linear|BSpline
|WindowedSinc>]
[--BackgroundFillValue
<std::string>]
[--rescaleIntensitiesOutputRange
<std::vector<int>>]
[--trimRescaledIntensities
<double>] [--rescaleIntensities]
[--outputUntransformedClippedVolume
<std::string>]
[--cutOutHeadInOutputVolume]
[--acLowerBound <double>]
[--otsuPercentileThreshold
<double>] [--mspQualityLevel <int>]
[--outputVerificationScript
<std::string>] [--outputMRML
<std::string>]
[--outputLandmarksPaired
<std::string>]
[--inputLandmarksPaired
<std::string>]
[--outputLandmarksInACPCAlignedSpac
e <std::string>]
[--outputLandmarksInInputSpace
<std::string>] [--outputTransform
<std::string>]
[--outputResampledVolume
<std::string>] [-o <std::string>]
[-i <std::string>]
[--inputEPCAModelTxt <std::string>]
[--inputEPCAModelMat <std::string>]
[--inputLLSModel <std::string>]
[--inputTemplateModel
<std::string>]
[--houghEyeDetectorMode <int>] [--]
[--version] [-h]

Where:

 --returnparameterfile <std::string>
   Filename in which to write simple return parameters (int, float,
   int-vector, etc.) as opposed to bulk return parameters (image,
   geometry, transform, measurement, table).

 --processinformationaddress <std::string>
   Address of a structure to store process information (progress, abort,
   etc.). (default: 0)

 --xml
   Produce xml description of command line arguments (default: 0)

 --echo
   Echo the command line arguments (default: 0)

 -w <int>,  --writedebuggingImagesLevel <int>
   This flag controls if debugging images are produced.  By default value
   of 0 is no images.  Anything greater than zero will be increasing
   level of debugging images. (default: 0)

 -r <std::string>,  --resultsDir <std::string>
   The directory for the debuging images to be written. (default: ./)

 --writeBranded2DImage <std::string>
   The filename for the 2D .png branded midline debugging image.  This
   will happen only in conjunction with requesting an outputVolume.

 -v,  --verbose
   Show more verbose output (default: 0)

 --debug
   Show internal debugging information. (default: 0)

 --rVN4 <double>
   Search radius for VN4 (default: -1)

 --rpc <double>
   Search radius for PC (default: -1)

 --rac <double>
   Search radius for AC (default: -1)

 --rmpj <double>
   Search radius for MPJ (default: -1)

 --forceHoughEyeDetectorReportFailure
   Flag indicates whether the Hough eye detector should report failure
   (default: 0)

 --inputLandmarksEMSP <std::string>
   The filename for the new subject-specific landmark definition file in
   the same format produced by Slicer3 (in .fcsv) with the landmarks in
   the estimated MSP aligned space to be loaded. The detector will only
   process landmarks not enlisted on the file.

 --forceRPPoint <std::vector<float>>
   Use this flag to manually specify the RP point from the original image
   on the command line. (default: 0)

 --forceVN4Point <std::vector<float>>
   Use this flag to manually specify the VN4 point from the original
   image on the command line. (default: 0)

 --forcePCPoint <std::vector<float>>
   Use this flag to manually specify the PC point from the original image
   on the command line. (default: 0)

 --forceACPoint <std::vector<float>>
   Use this flag to manually specify the AC point from the original image
   on the command line. (default: 0)

 --interpolationMode <NearestNeighbor|Linear|BSpline|WindowedSinc>
   Type of interpolation to be used when applying transform to moving
   volume.  Options are Linear, NearestNeighbor, BSpline, or WindowedSinc
   (default: Linear)

 --BackgroundFillValue <std::string>
   Fill the background of image with specified short int value. Enter
   number or use BIGNEG for a large negative number. (default: 0)

 --rescaleIntensitiesOutputRange <std::vector<int>>
   This pair of integers gives the lower and upper bounds on the signal
   portion of the output image.  Out-of-field voxels are taken from
   BackgroundFillValue. (default: 40,4000)

 --trimRescaledIntensities <double>
   Turn on clipping the rescaled image one-tailed on input.  Units of
   standard deviations above the mean.  Very large values are very
   permissive.  Non-positive value turns clipping off.  Defaults to
   removing 0.00001 of a normal tail above the mean. (default: 4.4172)

 --rescaleIntensities
   Flag to turn on rescaling image intensities on input. (default: 0)

 --outputUntransformedClippedVolume <std::string>
   Output image in which to store neck-clipped input image, with the use
   of --acLowerBound and maybe --cutOutHeadInUntransformedVolume.

 --cutOutHeadInOutputVolume
   Flag to cut out just the head tissue when producing an (un)transformed
   clipped volume. (default: 0)

 --acLowerBound <double>
   When generating a resampled output image, replace the image with the
   BackgroundFillValue everywhere below the plane This Far in physical
   units (millimeters) below (inferior to) the AC point (as found by the
   model.)  The oversize default was chosen to have no effect.  Based on
   visualizing a thousand masks in the IPIG study, we recommend a limit
   no smaller than 80.0 mm. (default: 1000)

 --otsuPercentileThreshold <double>
   This is a parameter to FindLargestForegroundFilledMask, which is
   employed when acLowerBound is set and an
   outputUntransformedClippedVolume is requested. (default: 0.01)

 --mspQualityLevel <int>
   Flag cotrols how agressive the MSP is estimated.  0=quick estimate (9
   seconds), 1=normal estimate (11 seconds), 2=great estimate (22
   seconds), 3=best estimate (58 seconds). (default: 2)

 --outputVerificationScript <std::string>
   The filename for the Slicer3 script that verifies the aligned
   landmarks against the aligned image file.  This will happen only in
   conjunction with saveOutputLandmarks and an outputVolume.

 --outputMRML <std::string>
   The filename for the new subject-specific scene definition file in the
   same format produced by Slicer3 (in .mrml format). Only the components
   that were specified by the user on command line would be generated.
   Compatible components include inputVolume, outputVolume,
   outputLandmarksInInputSpace, outputLandmarksInACPCAlignedSpace, and
   outputTransform.

 --outputLandmarksPaired <std::string>
   Paired use with inputLandmarks (in a format of .fcsv). It indicates
   the output acpc-aligned landmark list filename. The aligned landmarks
   are the landamrks that are defined in the file named inputLandmarks
   transformed by the acpc versor transform calculated by the
   constellation detector.

 --inputLandmarksPaired <std::string>
   Paired use with outputLandmarks. It indicates the input landmark list
   filename (in a format of .fcsv) which contains the landmarks to be
   trainsformed to acpc-aligned space. The alignment transform will be
   calculated form the base landmarks provided, and then the rest of the
   landmarks are directly transformed to the acpc-aligned space rather
   than to estimate one by one.

 --outputLandmarksInACPCAlignedSpace <std::string>
   The filename for the new subject-specific landmark definition file in
   the same format produced by Slicer3 (.fcsv) with the landmarks in the
   output image space (the detected RP, AC, PC, and VN4) in it to be
   written.

 --outputLandmarksInInputSpace <std::string>
   The filename for the new subject-specific landmark definition file in
   the same format produced by Slicer3 (.fcsv) with the landmarks in the
   original image space (the detected RP, AC, PC, and VN4) in it to be
   written.

 --outputTransform <std::string>
   The filename for the original space to ACPC alignment to be written
   (in .mat format).

 --outputResampledVolume <std::string>
   ACPC-aligned output image in a resampled unifor space.  Currently this
   is a 1mm, 256^3, Identity direction image.

 -o <std::string>,  --outputVolume <std::string>
   ACPC-aligned output image with the same voxels, but updated origin,
   and direction cosign so that the AC point would fall at the physical
   location (0.0,0.0,0.0), and the mid-sagital plane is the plane where
   physical L/R coordinate is 0.0.

 -i <std::string>,  --inputVolume <std::string>
   Input image in which to find ACPC points

 --inputEPCAModelTxt <std::string>
   Linear least squares model using EPCA matlab binary filename; the file
   includes light-weighted information as landmark names and their search
   radii (default: processingList.txt)

 --inputEPCAModelMat <std::string>
   Linear least squares model using EPCA matlab binary filename; the file
   includes matrices and vectors information (default: LME_EPCA.mat)

 --inputLLSModel <std::string>
   Linear least squares model filename in single txt (default:
   llsModel.txt)

 --inputTemplateModel <std::string>
   User-specified template model. (default: T1.mdl)

 --houghEyeDetectorMode <int>
   This flag controls the mode of Hough eye detector.  By default, value
   of 1 is for T1W images, while the value of 0 is for T2W and PD images.
   (default: 1)

 --,  --ignore_rest
   Ignores the rest of the labeled arguments following this flag.

 --version
   Displays version information and exits.

 -h,  --help
   Displays usage information and exits.

GUI Corrector

The GUI corrector is mainly used to correct bad/wrong landmarks generated by the automated constellation detector. It takes in a NIfTI MRI input image data and an optional input Slicer 3-compatible landmarks file, then it renders the image and provides a way for the user to manipulate landmarks. Finally, the user can save the corrected landmarks to a file which will be loaded by the constellation detector to recompute the result while overriding the enlisted landmarks on the saved file.

 BRAINSConstellationDetectorGUI
 [--returnparameterfile
 <std::string>]
 [--processinformationaddress
 <std::string>] [--xml] [--echo] [-o
 <std::string>] [-l <std::string>]
 [-i <std::string>] [--] [--version]
 [-h]

Where

 --returnparameterfile <std::string>
    Filename in which to write simple return parameters (int, float,
    int-vector, etc.) as opposed to bulk return parameters (image,
    geometry, transform, measurement, table).

 --processinformationaddress <std::string>
    Address of a structure to store process information (progress, abort,
    etc.). (default: 0)

 --xml
    Produce xml description of command line arguments (default: 0)

 --echo
    Echo the command line arguments (default: 0)

 -o <std::string>,  --outputLandmarks <std::string>
    The filename for the new subject-specific landmark definition file in
    the same format produced by Slicer3 (in a .fcsv format) with the
    landmarks in the original image space in it to be written.

 -l <std::string>,  --inputLandmarks <std::string>
    The filename for the new subject-specific landmark definition file in
    the same format produced by Slicer3 (in a .fcsv format) with the
    landmarks in the original image space in it to be read in.

 -i <std::string>,  --inputVolume <std::string>
    The filename of the input NIfTI volume to view/manipulate

 --,  --ignore_rest
    Ignores the rest of the labeled arguments following this flag.

 --version
    Displays version information and exits.

 -h,  --help
    Displays usage information and exits.