From Slicer Wiki
Jump to: navigation, search
Home < Documentation < 4.4 < Modules < dtiprocess

For the latest Slicer documentation, visit the read-the-docs.



User Interface
Output - Fractional Anisotropy (FA)
Output - Mean Diffusivity (MD)

General Information

Module Type & Category

Type: CLI

Category: Diffusion.NIRALPipeline

Authors, Collaborators & Contact

Author: Casey Goodlett


Contact: name, email

Module Description

Program title DTIProcess
Program description dtiprocess is a tool that handles tensor fields. It takes as an input a tensor field in nrrd format.\nIt can generate diffusion scalar properties out of the tensor field such as : FA (--fa_output), Gradient FA image (--fa_gradient_output), color FA (--color_fa_output), MD (--md_output), Frobenius norm (--frobenius_norm_output), lbd1, lbd2, lbd3 (--lambda{1,2,3}_output), binary map of voxel where if any of the eigenvalue is negative, the voxel is set to 1 (--negative_eigenvector_output) \nIt also creates 4D images out of the tensor field such as: Highest eigenvector map (highest eigenvector at each voxel) (--principal_eigenvector_output)\n Masking capabilities: For any of the processing done with dtiprocess, it's possible to apply it on a masked region of the tensor field. You need to use the --mask option for any of the option to be applied on that tensor field sub-region only. If you want to save the masked tensor field use the option --outmask and specify the new masked tensor field file name.\ndtiprocess also allows a range of transformations on the tensor fields. The transformed tensor field file name is specified with the option --deformation_output. There are 3 resampling interpolation methods specified with the tag --interpolation followed by the type to use (nearestneighbor, linear, cubic) Then you have several transformations possible to apply:\n\t- Affine transformations using as an input \n\t- itk affine transformation file (based on the itkAffineTransform class)\n\t- Affine transformations using rview (details and download at There are 2 versions of rview both creating transformation files called dof files. The old version of rview outputs text files containing the transformation parameters. It can be read in with the --dof_file option. The new version outputs binary dof files. These dof files can be transformed into human readable file with the dof2mat tool which is part of the rview package. So you need to save the output of dof2mat into a text file which can then be used with the -- newdof_file option. Usage example: dof2mat mynewdoffile.dof >> mynewdoffile.txt \t dtiprocess --dti_image mytensorfield.nhdr --newdof_file mynewdoffile.txt --rot_output myaffinetensorfield.nhdr \nNon linear transformations as an input: The default transformation file type is d-field (displacement field) in nrrd format. The option to use is --forward with the name of the file. If the transformation file is a h-field you have to add the option --hField.
Program version 1.0.1
Program documentation-url


Quick Tour of Features and Use

A list panels in the interface, their features, what they mean, and how to use them.

  • Input DTI:
    • DTI Image [----dti_image] : DTI tensor volume
  • Diffusion Tensor Measurements:
    • FA Output [----fa_output] [-f]: Fractional Anisotropy output file
    • MD Output [----md_output] [-m]: Mean Diffusivity output file
    • Gradients scale [----sigma] [-s]: Scale of gradients Default value: 2.0
    • FA Gradient Output [----fa_gradient_output] : Fractional Anisotropy Gradient output file
    • FA Gradient Magnitude Output [----fa_gradmag_output] : Fractional Anisotropy Gradient Magnitude output file
    • Color FA Output [----color_fa_output] : Color Fractional Anisotropy output file
    • Principal Eigenvectors Output [----principal_eigenvector_output] : Principal Eigenvectors Output
    • Negative Eigenvectors Output [----negative_eigenvector_output] : Negative Eigenvectors Output: create a binary image where if any of the eigen value is below zero, the voxel is set to 1, otherwise 0.
    • Frobenius Norm Output [----frobenius_norm_output] : Frobenius Norm Output
    • Lambda 1 (AD) [----lambda1_output] : Axial Diffusivity - Lambda 1 (largest eigenvalue) output
    • Lambda 2 [----lambda2_output] : Lambda 2 (middle eigenvalue) output
    • Lambda 3 [----lambda3_output] : Lambda 3 (smallest eigenvalue) output
    • RD [----RD_output] : RD (Radial Diffusivity 1/2*(lambda2+lambda3)) output
  • Affine Transformation:
    • Output Transformed Tensor [----rot_output] [-r]: Rotated tensor output file. Must also specify the dof file.
    • Affine Transform (ITK) [----affineitk_file] : Transformation file for affine transformation. ITK format.
    • Transform File (dof) [----dof_file] [-d]: Transformation file for affine transformation. This can be ITK format (or the outdated RView).
    • Transform File (new dof) [----newdof_file] : Transformation file for affine transformation. RView NEW format. (txt file output of dof2mat)
  • Masking:
    • Mask [----mask] : Mask tensors. Specify --outmask if you want to save the masked tensor field, otherwise the mask is applied just for the current processing
    • Masked tensor [----outmask] : Name of the masked tensor field.
  • Deformation field:
    • H Fields [----hField] : forward and inverse transformations are h-fields instead of displacement fields Default value: 0
    • Forward Transformation [----forward] [-F]: Forward transformation. Assumed to be a deformation field in world coordinates, unless the --h-field option is specified.
    • Deformed tensor image [----deformation_output] [-w]: Warped tensor field based on a deformation field. This option requires the --forward,-F transformation to be specified.
    • Interpolation [----interpolation] [-i]: Interpolation type (nearestneighbor, linear, cubic) Default value: linear
    • Reorientation [----reorientation] : Reorientation type (fs, ppd) Default value: fs
  • Options:
    • Scalar Float [----scalar_float] : Write scalar [FA,MD] as unscaled float (with their actual values, otherwise scaled by 10 000). Also causes FA to be unscaled [0..1]. Default value: 0
    • DTI double (do not use in 3D Slicer) [----DTI_double] : Tensor components are saved as doubles (cannot be visualized in Slicer) Default value: false
    • Verbose [----verbose] [-v]: produce verbose output Default value: 0
User Interface

More Information

More information is available on the project webpage:



Publications related to this module go here. Links to pdfs would be useful.