nipype.interfaces.niftyreg.reg module

The reg module provides classes for interfacing with the niftyreg registration command line tools.

The interfaces were written to work with niftyreg version 1.5.10

RegAladin

Link to code

Bases: NiftyRegCommand

Wrapped executable: reg_aladin.

Interface for executable reg_aladin from NiftyReg platform.

Block Matching algorithm for symmetric global registration. Based on Modat et al., “Global image registration using asymmetric block-matching approach” J. Med. Img. 1(2) 024003, 2014, doi: 10.1117/1.JMI.1.2.024003

Source code

Examples

>>> from nipype.interfaces import niftyreg
>>> node = niftyreg.RegAladin()
>>> node.inputs.ref_file = 'im1.nii'
>>> node.inputs.flo_file = 'im2.nii'
>>> node.inputs.rmask_file = 'mask.nii'
>>> node.inputs.omp_core_val = 4
>>> node.cmdline
'reg_aladin -aff im2_aff.txt -flo im2.nii -omp 4 -ref im1.nii -res im2_res.nii.gz -rmask mask.nii'
Mandatory Inputs:
  • flo_file (a pathlike object or string representing an existing file) – The input floating/source image. Maps to a command-line argument: -flo %s.

  • ref_file (a pathlike object or string representing an existing file) – The input reference/target image. Maps to a command-line argument: -ref %s.

Optional Inputs:
  • aff_direct_flag (a boolean) – Directly optimise the affine parameters. Maps to a command-line argument: -affDirect.

  • aff_file (a pathlike object or string representing a file) – The output affine matrix file. Maps to a command-line argument: -aff %s.

  • args (a string) – Additional parameters to the command. Maps to a command-line argument: %s.

  • cog_flag (a boolean) – Use the masks centre of mass to initialise the transformation. Maps to a command-line argument: -cog.

  • environ (a dictionary with keys which are a bytes or None or a value of class ‘str’ and with values which are a bytes or None or a value of class ‘str’) – Environment variables. (Nipype default value: {})

  • flo_low_val (a float) – Lower threshold value on floating image. Maps to a command-line argument: -floLowThr %f.

  • flo_up_val (a float) – Upper threshold value on floating image. Maps to a command-line argument: -floUpThr %f.

  • fmask_file (a pathlike object or string representing an existing file) – The input floating mask. Maps to a command-line argument: -fmask %s.

  • gpuid_val (an integer) – Device to use id. Maps to a command-line argument: -gpuid %i.

  • i_val (an integer >= 0) – Percent of inlier blocks. Maps to a command-line argument: -pi %d.

  • in_aff_file (a pathlike object or string representing an existing file) – The input affine transformation. Maps to a command-line argument: -inaff %s.

  • ln_val (an integer >= 0) – Number of resolution levels to create. Maps to a command-line argument: -ln %d.

  • lp_val (an integer >= 0) – Number of resolution levels to perform. Maps to a command-line argument: -lp %d.

  • maxit_val (an integer >= 0) – Maximum number of iterations. Maps to a command-line argument: -maxit %d.

  • nac_flag (a boolean) – Use nifti header to initialise transformation. Maps to a command-line argument: -nac.

  • nosym_flag (a boolean) – Turn off symmetric registration. Maps to a command-line argument: -noSym.

  • omp_core_val (an integer) – Number of openmp thread to use. Maps to a command-line argument: -omp %i. (Nipype default value: 1)

  • platform_val (an integer) – Platform index. Maps to a command-line argument: -platf %i.

  • ref_low_val (a float) – Lower threshold value on reference image. Maps to a command-line argument: -refLowThr %f.

  • ref_up_val (a float) – Upper threshold value on reference image. Maps to a command-line argument: -refUpThr %f.

  • res_file (a pathlike object or string representing a file) – The affine transformed floating image. Maps to a command-line argument: -res %s.

  • rig_only_flag (a boolean) – Do only a rigid registration. Maps to a command-line argument: -rigOnly.

  • rmask_file (a pathlike object or string representing an existing file) – The input reference mask. Maps to a command-line argument: -rmask %s.

  • smoo_f_val (a float) – Amount of smoothing to apply to floating image. Maps to a command-line argument: -smooF %f.

  • smoo_r_val (a float) – Amount of smoothing to apply to reference image. Maps to a command-line argument: -smooR %f.

  • v_val (an integer >= 0) – Percent of blocks that are active. Maps to a command-line argument: -pv %d.

  • verbosity_off_flag (a boolean) – Turn off verbose output. Maps to a command-line argument: -voff.

Outputs:
  • aff_file (a pathlike object or string representing a file) – The output affine file.

  • avg_output (a string) – Output string in the format for reg_average.

  • res_file (a pathlike object or string representing a file) – The output transformed image.

RegF3D

Link to code

Bases: NiftyRegCommand

Wrapped executable: reg_f3d.

Interface for executable reg_f3d from NiftyReg platform.

Fast Free-Form Deformation (F3D) algorithm for non-rigid registration. Initially based on Modat et al., “Fast Free-Form Deformation using graphics processing units”, CMPB, 2010

Source code

Examples

>>> from nipype.interfaces import niftyreg
>>> node = niftyreg.RegF3D()
>>> node.inputs.ref_file = 'im1.nii'
>>> node.inputs.flo_file = 'im2.nii'
>>> node.inputs.rmask_file = 'mask.nii'
>>> node.inputs.omp_core_val = 4
>>> node.cmdline
'reg_f3d -cpp im2_cpp.nii.gz -flo im2.nii -omp 4 -ref im1.nii -res im2_res.nii.gz -rmask mask.nii'
Mandatory Inputs:
  • flo_file (a pathlike object or string representing an existing file) – The input floating/source image. Maps to a command-line argument: -flo %s.

  • ref_file (a pathlike object or string representing an existing file) – The input reference/target image. Maps to a command-line argument: -ref %s.

Optional Inputs:
  • aff_file (a pathlike object or string representing an existing file) – The input affine transformation file. Maps to a command-line argument: -aff %s.

  • amc_flag (a boolean) – Use additive NMI. Maps to a command-line argument: -amc.

  • args (a string) – Additional parameters to the command. Maps to a command-line argument: %s.

  • be_val (a float) – Bending energy value. Maps to a command-line argument: -be %f.

  • cpp_file (a pathlike object or string representing a file) – The output CPP file. Maps to a command-line argument: -cpp %s.

  • environ (a dictionary with keys which are a bytes or None or a value of class ‘str’ and with values which are a bytes or None or a value of class ‘str’) – Environment variables. (Nipype default value: {})

  • fbn2_val (a tuple of the form: (an integer >= 0, an integer >= 0)) – Number of bins in the histogram for reference image for given time point. Maps to a command-line argument: -fbn %d %d.

  • fbn_val (an integer >= 0) – Number of bins in the histogram for reference image. Maps to a command-line argument: --fbn %d.

  • flo_smooth_val (a float) – Smoothing kernel width for floating image. Maps to a command-line argument: -smooF %f.

  • flwth2_thr_val (a tuple of the form: (an integer >= 0, a float)) – Lower threshold for floating image at the specified time point. Maps to a command-line argument: -fLwTh %d %f.

  • flwth_thr_val (a float) – Lower threshold for floating image. Maps to a command-line argument: --fLwTh %f.

  • fmask_file (a pathlike object or string representing an existing file) – Floating image mask. Maps to a command-line argument: -fmask %s.

  • fupth2_thr_val (a tuple of the form: (an integer >= 0, a float)) – Upper threshold for floating image at the specified time point. Maps to a command-line argument: -fUpTh %d %f.

  • fupth_thr_val (a float) – Upper threshold for floating image. Maps to a command-line argument: --fUpTh %f.

  • incpp_file (a pathlike object or string representing an existing file) – The input cpp transformation file. Maps to a command-line argument: -incpp %s.

  • jl_val (a float) – Log of jacobian of deformation penalty value. Maps to a command-line argument: -jl %f.

  • kld2_flag (an integer >= 0) – Use KL divergence as the similarity measure for a given time point. Maps to a command-line argument: -kld %d.

  • kld_flag (a boolean) – Use KL divergence as the similarity measure. Maps to a command-line argument: --kld.

  • le_val (a float) – Linear elasticity penalty term. Maps to a command-line argument: -le %f.

  • ln_val (an integer >= 0) – Number of resolution levels to create. Maps to a command-line argument: -ln %d.

  • lncc2_val (a tuple of the form: (an integer >= 0, a float)) – SD of the Gaussian for computing LNCC for a given time point. Maps to a command-line argument: -lncc %d %f.

  • lncc_val (a float) – SD of the Gaussian for computing LNCC. Maps to a command-line argument: --lncc %f.

  • lp_val (an integer >= 0) – Number of resolution levels to perform. Maps to a command-line argument: -lp %d.

  • maxit_val (an integer >= 0) – Maximum number of iterations per level. Maps to a command-line argument: -maxit %d.

  • nmi_flag (a boolean) – Use NMI even when other options are specified. Maps to a command-line argument: --nmi.

  • no_app_jl_flag (a boolean) – Do not approximate the log of jacobian penalty at control points only. Maps to a command-line argument: -noAppJL.

  • noconj_flag (a boolean) – Use simple GD optimization. Maps to a command-line argument: -noConj.

  • nopy_flag (a boolean) – Do not use the multiresolution approach. Maps to a command-line argument: -nopy.

  • nox_flag (a boolean) – Don’t optimise in x direction. Maps to a command-line argument: -nox.

  • noy_flag (a boolean) – Don’t optimise in y direction. Maps to a command-line argument: -noy.

  • noz_flag (a boolean) – Don’t optimise in z direction. Maps to a command-line argument: -noz.

  • omp_core_val (an integer) – Number of openmp thread to use. Maps to a command-line argument: -omp %i. (Nipype default value: 1)

  • pad_val (a float) – Padding value. Maps to a command-line argument: -pad %f.

  • pert_val (an integer >= 0) – Add perturbation steps after each optimization step. Maps to a command-line argument: -pert %d.

  • rbn2_val (a tuple of the form: (an integer >= 0, an integer >= 0)) – Number of bins in the histogram for reference image for given time point. Maps to a command-line argument: -rbn %d %d.

  • rbn_val (an integer >= 0) – Number of bins in the histogram for reference image. Maps to a command-line argument: --rbn %d.

  • ref_smooth_val (a float) – Smoothing kernel width for reference image. Maps to a command-line argument: -smooR %f.

  • res_file (a pathlike object or string representing a file) – The output resampled image. Maps to a command-line argument: -res %s.

  • rlwth2_thr_val (a tuple of the form: (an integer >= 0, a float)) – Lower threshold for reference image at the specified time point. Maps to a command-line argument: -rLwTh %d %f.

  • rlwth_thr_val (a float) – Lower threshold for reference image. Maps to a command-line argument: --rLwTh %f.

  • rmask_file (a pathlike object or string representing an existing file) – Reference image mask. Maps to a command-line argument: -rmask %s.

  • rupth2_thr_val (a tuple of the form: (an integer >= 0, a float)) – Upper threshold for reference image at the specified time point. Maps to a command-line argument: -rUpTh %d %f.

  • rupth_thr_val (a float) – Upper threshold for reference image. Maps to a command-line argument: --rUpTh %f.

  • smooth_grad_val (a float) – Kernel width for smoothing the metric gradient. Maps to a command-line argument: -smoothGrad %f.

  • ssd2_flag (an integer >= 0) – Use SSD as the similarity measure for a given time point. Maps to a command-line argument: -ssd %d.

  • ssd_flag (a boolean) – Use SSD as the similarity measure. Maps to a command-line argument: --ssd.

  • sx_val (a float) – Final grid spacing along the x axes. Maps to a command-line argument: -sx %f.

  • sy_val (a float) – Final grid spacing along the y axes. Maps to a command-line argument: -sy %f.

  • sz_val (a float) – Final grid spacing along the z axes. Maps to a command-line argument: -sz %f.

  • vel_flag (a boolean) – Use velocity field integration. Maps to a command-line argument: -vel.

  • verbosity_off_flag (a boolean) – Turn off verbose output. Maps to a command-line argument: -voff.

Outputs:
  • avg_output (a string) – Output string in the format for reg_average.

  • cpp_file (a pathlike object or string representing a file) – The output CPP file.

  • invcpp_file (a pathlike object or string representing a file) – The output inverse CPP file.

  • invres_file (a pathlike object or string representing a file) – The output inverse res file.

  • res_file (a pathlike object or string representing a file) – The output resampled image.