Postprocess module

PyMIALSRTK postprocessing functions.

It encompasses a High Resolution mask refinement and an N4 global bias field correction.

BinarizeImage

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Runs the MIAL SRTK mask image module.

Example

>>> from pymialsrtk.interfaces.postprocess import BinarizeImage
>>> maskImg = MialsrtkMaskImage()
>>> maskImg.inputs.input_image = 'input_image.nii.gz'
Mandatory Inputs

input_image (a pathlike object or string representing a file) – Input image filename to be binarized.

Outputs

output_srmask (a pathlike object or string representing a file) – Image mask (binarized input).

ConcatenateImageMetrics

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Concatenate metrics CSV files with the metrics computed on each labelmap.

Example

>>> from pymialsrtk.interfaces.postprocess import ConcatenateImageMetrics
>>> concat_metrics = ConcatenateImageMetrics()
>>> concat_metrics.inputs.input_metrics = ['sub-01_acq-haste_run-1_paramset_1_metrics.csv', 'sub-01_acq-haste_run-1_paramset2_metrics.csv']
concat_metrics.inputs.input_metrics_labels = ['sub-01_acq-haste_run-1_paramset_1_metrics_labels.csv', 'sub-01_acq-haste_run-1_paramset2_metrics_labels.csv']
>>> concat_metrics.run()
Optional Inputs

  • input_metrics (a list of items which are a pathlike object or string representing a file)

  • input_metrics_labels (a list of items which are a pathlike object or string representing a file)

Outputs

  • output_csv (a pathlike object or string representing a file)

  • output_csv_labels (a pathlike object or string representing a file)

FilenamesGeneration

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Generates final filenames from outputs of super-resolution reconstruction.

Example

>>> from pymialsrtk.interfaces.postprocess import FilenamesGeneration
>>> filenamesGen = FilenamesGeneration()
>>> filenamesGen.inputs.sub_ses = 'sub-01'
>>> filenamesGen.inputs.stacks_order = [3,1,4]
>>> filenamesGen.inputs.sr_id = 3
>>> filenamesGen.inputs.use_manual_masks = False
>>> filenamesGen.run()
Mandatory Inputs

  • multi_parameters (a boolean) – Whether multiple SR were reconstructed.

  • run_type (a string) – Type of run (preproc or sr).

  • sr_id (an integer) – Super-Resolution id.

  • stacks_order (a list of items which are any value) – List of stack run-id that specify the order of the stacks.

  • sub_ses (a string) – Subject and session BIDS identifier to construct output filename.

  • use_manual_masks (a boolean) – Whether masks were computed or manually performed.

Optional Inputs

TV_params (a list of items which are any value) – List iterables TV parameters processed.

Outputs

substitutions (a list of items which are any value) – Output correspondance between old and new filenames.

FilenamesGeneration.m_substitutions = []

ImageMetrics

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Compute various image metrics on a SR reconstructed image compared to a ground truth.

Example

>>> from pymialsrtk.interfaces.postprocess import ImageMetrics
>>> compute_metrics = ImageMetrics()
>>> compute_metrics.inputs.input_image = 'sub-01_acq-haste_rec-SR_id-1_T2w.nii.gz'
>>> compute_metrics.inputs.input_ref_image = 'sub-01_acq-haste_desc-GT_T2w.nii.gz'
>>> compute_metrics.inputs.input_ref_mask = 'sub-01_acq-haste_desc-GT_T2w_mask.nii.gz'
>>> compute_metrics.inputs.input_TV_parameters = {'in_loop': '10', 'in_deltat': '0.01', 'in_lambda': '2.5', 'in_bregman_loop': '3', 'in_iter': '50', 'in_step_scale': '1', 'in_gamma': '1', 'in_inner_thresh': '1e-05', 'in_outer_thresh': '1e-06'}
>>> concat_metrics.run()
Mandatory Inputs

  • input_TV_parameters (a dictionary with keys which are any value and with values which are any value)

  • input_image (a pathlike object or string representing a file) – Input image filename.

  • input_ref_image (a pathlike object or string representing a file) – Input reference image filename.

  • input_ref_mask (a pathlike object or string representing a file) – Input reference mask filename.

Optional Inputs

  • input_ref_labelmap (a pathlike object or string representing a file) – Input reference labelmap filename.

  • mask_input (a boolean) – Whether the image and reference should be masked when computing the metrics. (Nipype default value: True)

  • normalize_input (a boolean) – Whether the image and reference should be individually normalized before passing them into the metrics. (Nipype default value: True)

Outputs

  • output_metrics (a pathlike object or string representing a file) – Output CSV.

  • output_metrics_labels (a pathlike object or string representing a file) – Output per-label CSV.

ImageMetrics.norm_data(data)[source]

Normalize data into 0-1 range

MergeMajorityVote

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Perform majority voting to merge a list of label-wise labelmaps.

Example

>>> from pymialsrtk.interfaces.postprocess import MergeMajorityVote
>>> merge_labels = MergeMajorityVote()
>>> merge_labels.inputs.input_images = ['sub-01_acq-haste_run-1_labels_1.nii.gz', 'sub-01_acq-haste_run-1_labels_2.nii.gz', 'sub-01_acq-haste_run-1_labels_3.nii.gz']
>>> merge_labels.run()
Mandatory Inputs

input_images (a list of items which are a pathlike object or string representing a file) – Inputs label-wise labelmaps to be merged.

Outputs

output_image (a pathlike object or string representing a file) – Output label map.

MialsrtkN4BiasFieldCorrection

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Runs the MIAL SRTK slice by slice N4 bias field correction module.

This tools implements the method proposed by Tustison et al. [1]_ slice by slice.

References

1

Tustison et al.; Medical Imaging, IEEE Transactions, 2010. (link to paper)

Example

>>> from pymialsrtk.interfaces.postprocess import MialsrtkSliceBySliceN4BiasFieldCorrection
>>> N4biasFieldCorr = MialsrtkSliceBySliceN4BiasFieldCorrection()
>>> N4biasFieldCorr.inputs.input_image = 'sub-01_acq-haste_run-1_SR.nii.gz'
>>> N4biasFieldCorr.inputs.input_mask = 'sub-01_acq-haste_run-1_mask.nii.gz'
>>> N4biasFieldCorr.run()
Mandatory Inputs

input_image (a pathlike object or string representing a file) – Input image filename to be normalized.

Optional Inputs

  • input_mask (a pathlike object or string representing a file) – Input mask filename.

  • out_fld_postfix (a string) – (Nipype default value: _gbcorrfield)

  • out_im_postfix (a string) – (Nipype default value: _gbcorr)

  • verbose (a boolean) – Enable verbosity.

Outputs

  • output_field (a pathlike object or string representing a file) – Output bias field extracted from input image.

  • output_image (a pathlike object or string representing a file) – Output corrected image.

MialsrtkRefineHRMaskByIntersection

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Runs the MIALSRTK mask refinement module.

It uses the Simultaneous Truth And Performance Level Estimate (STAPLE) by Warfield et al. [1]_.

References

1

Warfield et al.; Medical Imaging, IEEE Transactions, 2004. (link to paper)

Example

>>> from pymialsrtk.interfaces.postprocess import MialsrtkRefineHRMaskByIntersection
>>> refMask = MialsrtkRefineHRMaskByIntersection()
>>> refMask.inputs.input_images = ['sub-01_acq-haste_run-1_T2w.nii.gz','sub-01_acq-haste_run-2_T2w.nii.gz']
>>> refMask.inputs.input_masks = ['sub-01_acq-haste_run-1_mask.nii.gz','sub-01_acq-haste_run-2_mask.nii.gz']
>>> refMask.inputs.input_transforms = ['sub-01_acq-haste_run-1_transform.txt','sub-01_acq-haste_run-2_transform.nii.gz']
>>> refMask.inputs.input_sr = 'sr_image.nii.gz'
>>> refMask.run()
Mandatory Inputs

input_sr (a pathlike object or string representing a file) – SR image filename.

Optional Inputs

  • in_use_staple (a boolean) – Use STAPLE for voting (default is True). If False, Majority voting is used instead. (Nipype default value: True)

  • input_images (a list of items which are a pathlike object or string representing a file) – Image filenames used in SR reconstruction.

  • input_masks (a list of items which are a pathlike object or string representing a file) – Mask filenames.

  • input_rad_dilatation (an integer) – Radius of the structuring element (ball). (Nipype default value: 1)

  • input_transforms (a list of items which are a pathlike object or string representing a file) – Transformation filenames.

  • out_lrmask_postfix (a string) – Suffix to be added to the Low resolution input_masks. (Nipype default value: _LRmask)

  • out_srmask_postfix (a string) – Suffix to be added to the SR reconstruction filename to construct output SR mask filename. (Nipype default value: _srMask)

  • verbose (a boolean) – Enable verbosity.

Outputs

  • output_lrmasks (a list of items which are a pathlike object or string representing a file) – Output low-resolution reconstruction refined masks.

  • output_srmask (a pathlike object or string representing a file) – Output super-resolution reconstruction refined mask.

ReportGeneration

Link to code

Bases: nipype.interfaces.base.core.BaseInterface

Generate a report summarizing the outputs of mialsrtk.

This comprises:

  • Details of which parameters were run.

  • A visualization of the SR-reconstructed image

  • A visualization of the computational graph.

Example

>>> from pymialsrtk.interfaces.postprocess import ReportGeneration
>>> report_gen = ReportGeneration()
>>> report_gen.inputs.subject = "sub-01"
>>> report_gen.inputs.session = "ses-01"
>>> report_gen.inputs.stacks = [2,3,5,1]
>>> report_gen.inputs.stacks_order = [5,2,3,1]
>>> report_gen.inputs.sr_id = "1"
>>> report_gen.inputs.run_type = "sr"
>>> report_gen.inputs.output_dir = "report_dir/
>>> report_gen.inputs.run_start_time = ""
>>> report_gen.inputs.run_elapsed_time = ""
>>> report_gen.inputs.do_refine_hr_mask = True
>>> report_gen.inputs.do_nlm_denoising = True
>>> report_gen.inputs.skip_stacks_ordering = False
>>> report_gen.inputs.skip_svr = False
>>> report_gen.inputs.masks_derivatives_dir = "derivatives/masks"
>>> report_gen.inputs.do_reconstruct_labels = False
>>> report_gen.inputs.do_anat_orientation = True
>>> report_gen.inputs.do_srr_assessment = False
>>> report_gen.inputs.openmp_number_of_cores = 3
>>> report_gen.inputs.nipype_number_of_cores = 1
>>> report_gen.inputs.input_sr = "outputs/sub-01_ses-01_res-SR_T2w.nii.gz"
>>> report_gen.inputs.input_json_path = "outputs/sub-01_ses-01_res-SR_T2w.json"
>>> report_gen.run()
Mandatory Inputs

  • do_anat_orientation (a boolean)

  • do_multi_parameters (a boolean)

  • do_nlm_denoising (a boolean)

  • do_reconstruct_labels (a boolean)

  • do_refine_hr_mask (a boolean)

  • do_srr_assessment (a boolean)

  • input_json_path (a pathlike object or string representing a file)

  • input_sr (a pathlike object or string representing a file)

  • masks_derivatives_dir (a string)

  • nipype_number_of_cores (an integer)

  • openmp_number_of_cores (an integer)

  • output_dir (a string)

  • run_elapsed_time (a float)

  • run_start_time (a float)

  • run_type (a string) – Type of run (preproc or sr).

  • session (a string) – Session BIDS identifier to construct output filename.

  • skip_stacks_ordering (a boolean)

  • skip_svr (a boolean)

  • sr_id (an integer) – Super-Resolution id.

  • stacks (a list of items which are any value) – List of stack run-id that specify the order of the stacks.

  • stacks_order (a list of items which are any value) – List of stack run-id that specify the order of the stacks.

  • subject (a string) – Subject BIDS identifier to construct output filename.

Optional Inputs

substitutions (a list of items which are any value) – Output correspondance between old and new filenames.

Outputs

report_html (a pathlike object or string representing a file)