brains:BRAINSClassify

From NITRC Wiki

Return to brains:MainPage

Summary

This program will perform the traditional BRAINS continuous tissue classification using generation of discriminant functions and applying the functions to the co-registered input images. Currently the program can support T1, T2, and PD image types. The algorithm had been optimized for multi-modal data (T1 and T2, or T1 and PD). It does work with single modal data as well, but has been tested less extensively. For example, venous blood will be classified as gray matter using a T1 only segmentation.

The discriminant functions are generated from the class plugs. The typical way to generate the plugs is based on the brains:BRAINSClassPlugs. However these could be generated by other classification procedures such as BRAINSABC or a k-Means classification.

The output is a continuous tissue classified image coded on an 8-bit scale. The scale is defined as follows:

• 10 - Pure CSF
• 130 - Pure grey matter
• 250 - Pure white matter

Partial volume of tissue types can exist between CSF and gray matter as well as gray matter and white matter. The relative concentration is based on the 8-bit value. For example, 190 represents a voxel that is 50% gray matter and 50% white matter. The other codes 0-9 and 251-255 are reserved for other discrete tissue types. Presently, the following two codes are used:

• 0 - Other tissue
• 1 - Venous blood

We have also reserved code 5 to represent white matter hyperintensities. This continuous tissue classified image can be converted to a discrete classification using brains:BrainsDiscreteClass. The continuous tissue classified image is used by our surface generation algorithm and neural network algorithms while the discrete image is used for volumetric measurements.

Authors

• Vincent A. Magnotta
• Greg Harris

Usage

  BRAINSClassify 
                  [--returnparameterfile <std::string>]
                  [--processinformationaddress
                  <std::string>] [--xml] [--echo]
                  [--excludeVolume <std::string>]
                  [--generateSeperateImages]
                  [--histogramEqualize] [--yz] [--xz]
                  [--xy] [--zz] [--yy] [--xx] [--z]
                  [--y] [--x] [--spatialTrim <float>]
                  [--grossTrim <float>]
                  [--classVolume
                   <std::vector<std::string>>] ... 
                  [--BrainVolume <std::string>]
                  [--bloodPlugs <std::string>]
                  [--csfPlugs <std::string>]
                  [--wmPlugs <std::string>]
                  [--gmPlugs <std::string>]
                  [--pdVolume <std::string>]
                  [--t2Volume <std::string>]
                  [--t1Volume <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)

--excludeVolume <std::string>
Mask used to define where T1 image is ignored

--generateSeperateImages
Generate a seperate floating point image for each class (default: 0)

--histogramEqualize
Perform histogram equalization on input images (default: 0)

--yz
Model bias field cross terms in YZ (default: 0)

--xz
Model bias field cross terms in XZ (default: 0)

--xy
Model bias field cross terms in XY (default: 0)

--zz
Model 2nd order bias field in Z (default: 0)

--yy
Model 2nd order bias field in Y (default: 0)

--xx
Model 2nd order bias field in X (default: 0)

--z
Model linear bias field in Z (default: 0)

--y
Model linear bias field in Y (default: 0)

--x
Model linear bias field in X (default: 0)

--spatialTrim <float>
Spatial Trim (default: 1)

--grossTrim <float>
Gross Trim (default: 1)

--classVolume <std::vector<std::string>> (accepted multiple times)
Continuous Tissue Classified Image

--BrainVolume <std::string>
Brain Mask used for Histogram Equalization

--bloodPlugs <std::string>
Plugs for Venous Blood

--csfPlugs <std::string>
Plugs for CSF

--wmPlugs <std::string>
Plugs for White Matter

--gmPlugs <std::string>
Plugs for Grey Matter

--pdVolume <std::string>
PD Volume

--t2Volume <std::string>
T2 Volume

--t1Volume <std::string>
T1 Volume

--, --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.

Figures

Gray Matter Plugs

White Matter Plugs

References

1. Harris G, Andreasen NC, Cizadlo T, Bailey JM, Bockholt HJ, Magnotta VA, Arndt S. Improving tissue classification in MRI: a three-dimensional multispectral discriminant analysis method with automated training class selection. J Comput Assist Tomogr. 1999 Jan-Feb;23(1):144-54.

Acknowledgements

This work was developed by the University of Iowa Departments of Radiology and Psychiatry. This software was supported in part of NIH/NINDS award NS050568.