cmp.stages.registration.registration module

Definition of config and stage classes for MRI co-registration

class cmp.stages.registration.registration.RegistrationConfig

Bases: traits.has_traits.HasTraits

Class used to store configuration parameters of a RegistrationStage instance.

pipeline

Pipeline type (Default: “Diffusion”)

Type

traits.Enum([“Diffusion”, “fMRI”])

registration_mode_trait

Choices of registration tools updated depending on the pipeline type. (Default: [‘FSL’, ‘ANTs’] if “Diffusion”, [‘FSL’, ‘BBregister (FS)’] if “fMRI”)

Type

traits.List([‘FSL’, ‘ANTs’, ‘BBregister (FS)’])

registration_mode

Registration tool used from the registration_mode_trait list (Default: ‘ANTs’)

Type

traits.Str

diffusion_imaging_model

Diffusion imaging model (‘DTI’ for instance)

Type

traits.Str

use_float_precision

Use ‘single’ instead of ‘double’ float representation to reduce memory usage of ANTs (Default: False)

Type

traits.Bool

ants_interpolation

Interpolation type used by ANTs that can be: ‘Linear’, ‘NearestNeighbor’, ‘CosineWindowedSinc’, ‘WelchWindowedSinc’, ‘HammingWindowedSinc’, ‘LanczosWindowedSinc’, ‘BSpline’, ‘MultiLabel’, or ‘Gaussian’ (Default: ‘Linear’)

Type

traits.Enum

ants_bspline_interpolation_parameters

ANTs BSpline interpolation parameters (Default: traits.Tuple(Int(3)))

Type

traits.Tuple

ants_gauss_interpolation_parameters

ANTs Gaussian interpolation parameters (Default: traits.Tuple(Float(5), Float(5)))

Type

traits.Tuple

ants_multilab_interpolation_parameters

ANTs Multi-label interpolation parameters (Default: traits.Tuple(Float(5), Float(5)))

Type

traits.Tuple

ants_lower_quantile

ANTs lower quantile (Default: 0.005)

Type

traits.Float

ants_upper_quantile

ANTs upper quantile (Default: 0.995)

Type

traits.Float

ants_convergence_thresh

ANTs convergence threshold (Default: 1e-06)

Type

traits.Float

ants_convergence_winsize

ANTs convergence window size (Default: 10)

Type

traits.Int

ants_linear_gradient_step

ANTS linear gradient step size (Default: 0.1)

Type

traits.Float

ants_linear_cost

Metric used by ANTs linear registration phase that can be ‘CC’, ‘MeanSquares’, ‘Demons’, ‘GC’, ‘MI’, or ‘Mattes’ (Default: ‘MI’)

Type

traits.Enum

ants_linear_sampling_strategy

ANTS sampling strategy for the linear registration phase that can be ‘None’, ‘Regular’, or ‘Random’ (Default: ‘Regular’)

Type

traits.Enum

ants_linear_sampling_perc

Percentage used if random sampling strategy is employed in the linear registration phase (Default: 0.25)

Type

traits.Float

ants_perform_syn

(Default: True)

Type

traits.Bool

ants_nonlinear_gradient_step

(Default: 0.1)

Type

traits.Float

ants_nonlinear_cost

Metric used by ANTs nonlinear (SyN) registration phase that can be ‘CC’, ‘MeanSquares’, ‘Demons’, ‘GC’, ‘MI’, or ‘Mattes’ (Default: ‘CC’)

Type

traits.Enum

ants_nonlinear_update_field_variance

Weight to update field variance in ANTs nonlinear (SyN) registration phase (Default: 3.0)

Type

traits.Float

ants_nonlinear_total_field_variance

Weight to give to total field variance in ANTs nonlinear (SyN) registration phase (Default: 0.0)

Type

traits.Float

flirt_args

FLIRT extra arguments that will be append to the FSL FLIRT command (Default: None)

Type

traits.Str

uses_qform

FSL FLIRT uses qform (Default: True)

Type

traits.Bool

dof

Specify number of degree-of-freedom to FSL FLIRT (Default: 6)

Type

traits.Int

fsl_cost

Metric used by FSL registration that can be ‘mutualinfo’, ‘corratio’, ‘normcorr’, ‘normmi’, ‘leastsq’, or ‘labeldiff’ (Default: ‘normmi’)

Type

traits.Enum

no_search

Enable FSL FLIRT “no search” option (Default: True)

Type

traits.Bool

init

Initialization type of FSL registration: ‘spm’, ‘fsl’, or ‘header’ (Default: ‘smp’)

Type

traits.Enum(‘header’, [‘spm’, ‘fsl’, ‘header’])

contrast_type

Contrast type specified to BBRegister: ‘t1’, ‘t2’, or ‘dti’ (Default: ‘dti’)

Type

traits.Enum(‘dti’, [‘t1’, ‘t2’, ‘dti’])

apply_to_eroded_wm

Apply estimated transform to eroded white-matter mask (Default: True)

Type

traits.Bool

apply_to_eroded_csf

Apply estimated transform to eroded cortico spinal fluid mask (Default: True)

Type

traits.Bool

apply_to_eroded_brain

Apply estimated transform to eroded brain mask (Default: False)

Type

traits.Bool

See also

cmp.stages.registration.registration.RegistrationStage

class cmp.stages.registration.registration.RegistrationStage(pipeline_mode, fs_subjects_dir=None, fs_subject_id=None, bids_dir='', output_dir='')

Bases: cmp.stages.common.Stage

Class that represents the registration stage of both DiffusionPipeline and fMRIPipeline.

fs_subjects_dir

Freesurfer subjects directory (needed by BBRegister)

Type

traits.Directory

fs_subject_id

Freesurfer subject (being processed) directory (needed by BBRegister)

Type

traits.Str

create_workflow()

Create the workflow of the RegistrationStage

See also

cmp.pipelines.diffusion.diffusion.DiffusionPipeline, cmp.pipelines.functional.fMRI.fMRIPipeline, cmp.stages.registration.registration.RegistrationConfig

create_workflow(flow, inputnode, outputnode)

Create the stage worflow.

Parameters

• flow (nipype.pipeline.engine.Workflow) – The nipype.pipeline.engine.Workflow instance of either the Diffusion pipeline or the fMRI pipeline

• inputnode (nipype.interfaces.utility.IdentityInterface) – Identity interface describing the inputs of the stage

• outputnode (nipype.interfaces.utility.IdentityInterface) – Identity interface describing the outputs of the stage

define_inspect_outputs()

Update the `inspect_outputs’ class attribute.

It contains a dictionary of stage outputs with corresponding commands for visual inspection.

has_run()

Function that returns True if the stage has been run successfully.

Returns

Return type

True if the stage has been run successfully

cmp.stages.registration.registration.unicode2str(text)

Convert a unicode to a string using system’s encoding.

(This methods should be moved to utils in the upcoming future.)

Parameters

text (bytes) – Unicode bytes representation of a string

Returns

out_str – Output string

Return type

str