deep_folding.brainvisa package

Submodules

deep_folding.brainvisa.benchmark_generation_distmap module

Creating abnormal skeleton images

The aim of this script is to generate a benchmark of sulcal abnormalities. Abnormalities are defined as skeletons with one simple surface missing. This simple surface must be completely inside the bounding box of interest and include a minimum number of voxels (in order that the anomaly is big enough to be considered as abnormal).

class deep_folding.brainvisa.benchmark_generation_distmap.Benchmark(b_num, side, ss_size, sulci_list, saving_dir, data_dir='/neurospin/dico/data/bv_databases/human/hcp/hcp', bbox_dir='/neurospin/dico/data/deep_folding/current/bbox/', mask_dir='/neurospin/dico/data/deep_folding/current/mask/1mm/', mask=True, inpainting=False)

Bases: object

Generates benchmark of altered skeletons

delete_ss(sub)

Deletes one simple surface

Parameters:sub – int giving the subject
generate_foldlabel(sub)

Deletes one simple surface :param sub: int giving the subject

generate_skeleton(sub)

Generates a skeleton from a graph

Parameters:sub – int giving the subject
get_simple_surfaces(sub)
Selects simple surfaces of one subject that satisfy following
conditions: size in [given ss_size min, given ss_size max] and with the right number of voxels included in the mask
Parameters:sub – int giving the subject
save_file(sub)

Saves the modified skeleton

Parameters:sub – int giving the subject
save_lists(abnormality_test, givers, subjects_list)

Saves lists of modified subjects

Parameters:
  • abnormality_test – list of modified subjects
  • givers – in case of simple surface addition, list of subjects whose one of the simple surface has been added to another subject
  • subjects_list – list of subjects
deep_folding.brainvisa.benchmark_generation_distmap.generate(b_num, side, ss_size, sulci_list, bench_size=150, subjects_list=None, saving_dir='/neurospin/dico/lguillon/distmap/benchmark/test/', bbox_dir='/neurospin/dico/data/deep_folding/current/bbox/', inpainting=False)

Generates a benchmark

Parameters:
  • b_num – number of the benchmark to create
  • side – hemisphere side (either L for left, or R for right hemisphere)
  • sulci_list – list of sulcus names
deep_folding.brainvisa.benchmark_generation_distmap.get_sub_list(subjects_list)

Returns subjects list for which latered skelerons can be created Only right handed HCP subjects

deep_folding.brainvisa.benchmark_pipeline module

Full pipeline to create a benchmark of abnormalities.

class deep_folding.brainvisa.benchmark_pipeline.BenchmarkPipe(tgt_dir, sulcus, side, ss_size, mode, bench_size, resampling, bbox_dir, subjects_list)

Bases: object

Generates benchmark files of a given mode

crop_one_file(sub)

Normalizes, crops and flips if necessary, one file

Parameters:sub – string giving the subject ID
get_sub_list()

Get IDs of subjects to normalize, crop

launch_pipe()

Main API to create benchmark files

deep_folding.brainvisa.benchmark_pipeline.define_njobs()

Returns number of cpus used by main loop

deep_folding.brainvisa.benchmark_pipeline.main(argv)

Reads argument line and creates benchmark files

Parameters:argv – a list containing command line arguments
deep_folding.brainvisa.benchmark_pipeline.parse_args(argv)

Function parsing command-line arguments

Parameters:argv – a list containing command line arguments
Returns:a list with source directory names, full path sulcus: a string containing the sulcus to analyze number_subjects: number of subjects to analyze
Return type:src_dir

deep_folding.brainvisa.compute_bounding_box module

The aim of this script is to compute the bounding box for a given sulcus based on a manually labelled dataset

Bounding box corresponds to the biggest box that encompasses the given sulci on all subjects of the manually labelled dataset. It measures the bounding box in the MNI152 space.

class deep_folding.brainvisa.compute_bounding_box.BoundingBoxMax(src_dir='/neurospin/dico/data/bv_databases/human/manually_labeled/pclean/all', path_to_graph='t1mri/t1/default_analysis/folds/3.3/base2018_manual', bbox_dir='test/bbox/1mm', sulcus='F.C.M.ant.', new_sulcus=None, side='R', out_voxel_size=None)

Bases: object

Determines the maximum Bounding Box around given sulci

It is determined in the MNI ICBM152 nonlinear 2009c asymmetrical template http://www.bic.mni.mcgill.ca/~vfonov/icbm/2009/mni_icbm152_nlin_asym_09b_nifti.zip

compute(number_subjects=-1)

Main class program to compute the bounding box

Parameters:number_subjects – number_subjects to analyze
deep_folding.brainvisa.compute_bounding_box.box_ICBM2009c_to_aims_talairach(bbmin_mni152: <built-in function array>, bbmax_mni152: <built-in function array>) → tuple

Transform bbox coordinates from MNI152 to AIMS talairach referential

deep_folding.brainvisa.compute_bounding_box.compute_bounding_box(src_dir='/neurospin/dico/data/bv_databases/human/manually_labeled/pclean/all', bbox_dir='test/bbox/1mm', path_to_graph='t1mri/t1/default_analysis/folds/3.3/base2018_manual', sulcus='F.C.M.ant.', new_sulcus=None, side='R', number_subjects=-1, out_voxel_size=None)

Main program computing the box encompassing the sulcus in all subjects

The programm loops over all subjects and computes in MNI space the voxel coordinates of the box encompassing the sulci for all subjects

Parameters:
  • src_dir – list of strings -> directories of the supervised databases
  • bbox_dir – string giving target bbox directory path
  • path_to_graph – string giving relative path to manually labelled graph
  • side – hemisphere side (either ‘L’ for left, or ‘R’ for right)
  • sulcus – string giving the sulcus to analyze
  • new_sulcus – string giving a new sulcus name (optional)
  • number_subjects – integer giving the number of subjects to analyze, by default it is set to _ALL_SUBJECTS (-1)
  • out_voxel_size – float giving voxel size in mm
Returns:

tuple (bbmin_vox, bbmax_vox)
deep_folding.brainvisa.compute_bounding_box.compute_box_voxel(bbmin_mni152, bbmax_mni152, voxel_size_out)

Returns the coordinates of the box as voxels

Coordinates of the box in voxels are determined in the MNI referential

Parameters:
  • bbmin_mni152 – numpy array with coordinates of the upper right corner of the box (MNI152 space)
  • bbmax_mni152 – numpy array with coordinates of the lower left corner of the box (MNI152 space)
  • voxel_size – voxel size (in MNI referential or HCP normalized SPM space)
Returns:

tuple (bbmin_vox, bbmax_vox) with
bbmin_vox: numpy array with coordinates of the upper right corner

of the box (voxels in MNI space);

bbmax_vox: numpy array with coordinates of the lower left corner

of the box (voxels in MNI space)
deep_folding.brainvisa.compute_bounding_box.get_bounding_boxes(subjects, sulcus)

get bounding boxes of the chosen sulcus for all subjects.

Function that outputs bounding box for the listed sulci on a manually labeled dataset. Bounding box corresponds to the biggest box encountered in the manually labeled subjects in the MNI1 152 space. The bounding box is the smallest rectangular box that encompasses the sulcus.

Parameters:
  • subjects – list containing all subjects to be analyzed
  • sulcus – str -> sulcus to be analyzed
Returns:

list_bbmin: list containing the upper right vertex of the box

in the MNI 152 space;

list_bbmax: list containing the lower left vertex of the box

in the MNI 152 space
Return type:

tuple (list_bbmin, list_bbmax) with
deep_folding.brainvisa.compute_bounding_box.get_one_bounding_box(graph_filename, sulcus)

get bounding box of the chosen sulcus for one data graph in MNI 152

Function that outputs the bounding box for the listed sulci for this datagraph. The bounding box is the smallest rectangular box that encompasses the chosen sulcus. It is given in the MNI 152 referential.

Parameters:graph_filename – string being the name of graph file .arg to analyze: for example: ‘Lammon_base2018_manual.arg’
Returns:
numpy array giving the upper right vertex coordinates
of the box in the MNI 152 referential
bbox_max: numpy array fiving the lower left vertex coordinates
of the box in the MNI 152 referential
Return type:bbox_min
deep_folding.brainvisa.compute_bounding_box.parse_args(argv: list) → dict

Function parsing command-line arguments

Parameters:argv – a list containing command line arguments
Returns:a dictionary with all arugments as keys
Return type:params

deep_folding.brainvisa.compute_mask module

The aim of this script is to output bounding box and mask for a given sulcus based on a manually labelled dataset

Bounding box corresponds to the biggest box that encompasses the given sulci on all subjects of the manually labelled dataset. It measures the bounding box in the MNI152 space.

class deep_folding.brainvisa.compute_mask.MaskAroundSulcus(src_dir='/neurospin/dico/data/bv_databases/human/manually_labeled/pclean/all', path_to_graph='t1mri/t1/default_analysis/folds/3.3/base2018_manual', mask_dir='test/mask/1mm', sulcus='F.C.M.ant.', new_sulcus=None, side='R', out_voxel_size=None)

Bases: object

Determines the maximum Bounding Box around given sulci

It is determined in the MNI ICBM152 nonlinear 2009c asymmetrical template http://www.bic.mni.mcgill.ca/~vfonov/icbm/2009/mni_icbm152_nlin_asym_09b_nifti.zip

compute(number_subjects=-1)

Main class program to compute the bounding box

Parameters:number_subjects – number_subjects to analyze
deep_folding.brainvisa.compute_mask.compute_mask(src_dir='/neurospin/dico/data/bv_databases/human/manually_labeled/pclean/all', mask_dir='test/mask/1mm', path_to_graph='t1mri/t1/default_analysis/folds/3.3/base2018_manual', sulcus='F.C.M.ant.', new_sulcus=None, side='R', number_subjects=-1, out_voxel_size=None)

Main program computing the box encompassing the sulcus in all subjects

The programm loops over all subjects and computes in MNI space the voxel coordinates of the box encompassing the sulci for all subjects

Parameters:
  • src_dir – list of strings -> directories of the supervised databases
  • mask_dir – string giving target mask directory path
  • path_to_graph – string giving relative path to manually labelled graph
  • side – hemisphere side (either ‘L’ for left, or ‘R’ for right)
  • sulcus – string giving the sulcus to analyze
  • new_sulcus – string giving a new sulcus name (optional)
  • number_subjects – integer giving the number of subjects to analyze, by default it is set to _ALL_SUBJECTS (-1)
  • out_voxel_size – float giving voxel size
Returns:

aims volume containing the mask
deep_folding.brainvisa.compute_mask.increment_mask(subjects: list, mask: <function Volume at 0x7f32bb7240d0>, sulcus: str, voxel_size_out: tuple, sample_dir: str)

Increments mask for the chosen sulcus for all subjects

Parameters:subjects – list containing all subjects to be analyzed
deep_folding.brainvisa.compute_mask.increment_one_mask(graph_filename, mask, sulcus, voxel_size_out)

Increments self.mask of 1 where there is the sulcus

Parameters:graph_filename – string being the name of graph file .arg to analyze: for example: ‘Lammon_base2018_manual.arg’
deep_folding.brainvisa.compute_mask.initialize_mask(out_voxel_size: tuple) → <function Volume at 0x7f32bb7240d0>

Creates aims volume in MNI ICBM152 nonlinear 2009c asymmetrical template http://www.bic.mni.mcgill.ca/~vfonov/icbm/2009/mni_icbm152_nlin_asym_09b_nifti.zip

Parameters:output_voxel_size – tuple Output voxel size (default: None, no resampling)
Returns:
volume (aims.Volume_S16) filled with 0 in MNI2009 referential
and with requested voxel_size
Return type:mask
deep_folding.brainvisa.compute_mask.parse_args(argv: list) → dict

Function parsing command-line arguments

Parameters:argv – a list containing command line arguments
Returns:a dictionary with all arugments as keys
Return type:params
deep_folding.brainvisa.compute_mask.write_mask(mask: <function Volume at 0x7f32bb7240d0>, mask_file: str)

Writes mask on mask file

deep_folding.brainvisa.display_reconstructions module

The aim of this script is to display saved model outputs with Anatomist. Model outputs are stored as numpy arrays.

deep_folding.brainvisa.display_reconstructions.array_to_ana(ana_a, img, sub_id, phase, status, bucket, vs)

Transforms output tensor into volume readable by Anatomist and defines name of the volume displayed in Anatomist window. Returns volume displayable by Anatomist either on the skeleton mode, that can be viewed in consecutive 2D slices, or on 3D buckets.

deep_folding.brainvisa.display_reconstructions.main()

In the Anatomist window, for each model output, corresponding input will also be displayed at its left side. Number of columns and view (Sagittal, coronal, frontal) can be specified. (It’s better to choose an even number for number of columns to display)

deep_folding.brainvisa.generate_ICBM2009c_transforms module

Write transform files to ICBM 2009c from graph files

Typical usage

You can use this program by first entering in the brainvisa environment (here brainvisa 5.0.0 installed with singurity) and launching the script from the terminal: >>> bv bash >>> python generate_ICBM2009c_transforms.py

class deep_folding.brainvisa.generate_ICBM2009c_transforms.GraphGenerateTransform(src_dir, transform_dir, side, parallel, path_to_graph, bids)

Bases: object

Class to convert all graphs from a folder into skeletons

It contains all information to scan a dataset for graphs and writes skeletons and foldlabels into target directory

compute(number_subjects)

Loops over subjects to generate transforms to ICBM2009c from graphs.

generate_one_transform(subject: str)

Generates and writes ICBM2009c transform for one subject.

get_transform_filename(subject, graph_file)
deep_folding.brainvisa.generate_ICBM2009c_transforms.generate_ICBM2009c_transforms(src_dir='/neurospin/dico/data/bv_databases/human/not_labeled/hcp/hcp', transform_dir='test/datasets/hcp/transforms', path_to_graph='t1mri/BL/default_analysis/folds/3.1', side='R', bids=False, parallel=False, number_subjects=-1)

Generates skeletons from graphs

deep_folding.brainvisa.generate_ICBM2009c_transforms.parse_args(argv)

Parses command-line arguments

Parameters:argv – a list containing command line arguments
Returns:args

deep_folding.brainvisa.generate_crops module

Creating npy file from T1 MRI datas The aim of this script is to create dataset of cropped skeletons from MRIs saved in a .npy file. We read resampled skeleton files Several steps are required: crop and .npy generation

You can use this program by first entering in the brainvisa environment (here brainvisa 5.0.0 installed with singurity) and launching the script from the terminal: >>> bv bash >>> python3 generate_crops.py Alternatively, you can launch the script in the interactive terminal ipython: >>> %run generate_crops.py
class deep_folding.brainvisa.generate_crops.CropGenerator(src_dir='test/datasets/hcp/skeletons/1mm', crop_dir='test/datasets/hcp/crops/CINGULATE/mask/1mm', bbox_dir='test/bbox/1mm', mask_dir='test/mask/1mm', dilation=5.0, threshold=0.0, list_sulci='F.C.M.ant.', side='R', cropping_type='mask', combine_type=False, parallel=False, no_mask=False)

Bases: object

Generates cropped skeleton files and corresponding npy file

compute(number_subjects=-1)

Main API to create numpy files The programm loops over all subjects from the input (source) directory. :param number_subjects: integer giving the number of subjects to analyze,

by default it is set to _ALL_SUBJECTS (-1).
compute_bounding_box_or_mask(number_subjects)

Computes bounding box or mask :param number_subjects: integer giving the number of subjects to analyze,

by default it is set to _ALL_SUBJECTS (-1).
crop_files(number_subjects=-1)

Crop nii files The programm loops over all subjects from the input (source) directory. :param number_subjects: integer giving the number of subjects to analyze,

by default it is set to _ALL_SUBJECTS (-1).
crop_one_file(subject_id)

Crops one file :param subject_id: string giving the subject ID

class deep_folding.brainvisa.generate_crops.DistMapCropGenerator(src_dir='test/datasets/hcp/skeletons/1mm', crop_dir='test/datasets/hcp/crops/CINGULATE/mask/1mm', bbox_dir='test/bbox/1mm', mask_dir='test/mask/1mm', list_sulci='F.C.M.ant.', side='R', cropping_type='mask', combine_type=False, parallel=False, no_mask=False)

Bases: deep_folding.brainvisa.generate_crops.CropGenerator

Generates cropped skeleton files and corresponding npy file

class deep_folding.brainvisa.generate_crops.FoldLabelCropGenerator(src_dir='test/datasets/hcp/skeletons/1mm', crop_dir='test/datasets/hcp/crops/CINGULATE/mask/1mm', bbox_dir='test/bbox/1mm', mask_dir='test/mask/1mm', list_sulci='F.C.M.ant.', side='R', cropping_type='mask', combine_type=False, parallel=False, no_mask=False)

Bases: deep_folding.brainvisa.generate_crops.CropGenerator

Generates cropped skeleton files and corresponding npy file

class deep_folding.brainvisa.generate_crops.SkeletonCropGenerator(src_dir='test/datasets/hcp/skeletons/1mm', crop_dir='test/datasets/hcp/crops/CINGULATE/mask/1mm', bbox_dir='test/bbox/1mm', mask_dir='test/mask/1mm', list_sulci='F.C.M.ant.', side='R', cropping_type='mask', combine_type=False, parallel=False, no_mask=False)

Bases: deep_folding.brainvisa.generate_crops.CropGenerator

Generates cropped skeleton files and corresponding npy file

deep_folding.brainvisa.generate_crops.crop_bbox(file_src: str, file_cropped: str, bbmin: <built-in function array>, bbmax: <built-in function array>)

Crops according to bounding box

deep_folding.brainvisa.generate_crops.crop_mask(file_src, file_cropped, mask, bbmin, bbmax, side, no_mask=False)

Crops according to mask

deep_folding.brainvisa.generate_crops.filter_mask(mask: <function Volume at 0x7f32bb7240d0>)

Smooths the mask with Gaussian Filter

deep_folding.brainvisa.generate_crops.generate_crops(src_dir='test/datasets/hcp/skeletons/1mm', input_type='skeleton', crop_dir='test/datasets/hcp/crops/CINGULATE/mask/1mm', bbox_dir='test/bbox/1mm', mask_dir='test/mask/1mm', side='R', list_sulci='F.C.M.ant.', number_subjects=-1, cropping_type='mask', combine_type=False, parallel=False, no_mask=True)
deep_folding.brainvisa.generate_crops.parse_args(argv)

Function parsing command-line arguments :param argv: a list containing command line arguments

Returns:dictionary with keys: src_dir, tgt_dir, nb_subjects, list_sulci
Return type:params

deep_folding.brainvisa.generate_distmaps module

Write distance maps from skeleton files

Note: generation of distance maps should be performed for each resolution

Typical usage

You can use this program by first entering in the brainvisa environment (here brainvisa 5.0.0 installed with singurity) and launching the script from the terminal: >>> bv bash >>> python generate_distmap.py

class deep_folding.brainvisa.generate_distmaps.SkelConvert2DistMap(src_dir, distmaps_dir, side, parallel, resampled_skel=False)

Bases: object

Class to convert all skeletons from a folder into distance maps

It contains all information to scan a dataset for skeletons and writes distance maps into target directory

compute(number_subjects)

Loops over subjects and converts graphs into distmaps.

generate_one_distmap(subject: str)

Generates and writes distmap for one subject.

deep_folding.brainvisa.generate_distmaps.generate_distmaps(src_dir='test/datasets/hcp/skeletons/raw', distmaps_dir='test/datasets/hcp/distmaps/raw', side='R', parallel=False, resampled_skel=True, number_subjects=-1)

Generates distmaps from skeletons

deep_folding.brainvisa.generate_distmaps.parse_args(argv)

Parses command-line arguments

Parameters:argv – a list containing command line arguments
Returns:args

deep_folding.brainvisa.generate_foldlabels module

Write foldlabels from graph files

Typical usage

You can use this program by first entering in the brainvisa environment (here brainvisa 5.0.0 installed with singurity) and launching the script from the terminal: >>> bv bash >>> python generate_foldlabels.py

class deep_folding.brainvisa.generate_foldlabels.GraphConvert2FoldLabel(src_dir, foldlabel_dir, side, junction, parallel, path_to_graph, bids, qc_path)

Bases: object

Class to convert graph into foldlabel files

It contains all information to scan a dataset for graphs and writes foldlabels into target directory

compute(number_subjects)

Loops over subjects and converts graphs into skeletons.

generate_one_foldlabel(subject: str)

Generates and writes skeleton for one subject.

get_foldlabel_filename(subject, graph_file)
deep_folding.brainvisa.generate_foldlabels.generate_foldlabels(src_dir='/neurospin/dico/data/bv_databases/human/not_labeled/hcp/hcp', foldlabel_dir='test/datasets/hcp/foldlabels/raw', path_to_graph='t1mri/BL/default_analysis/folds/3.1', side='R', junction='thin', bids=False, parallel=False, number_subjects=-1, qc_path='')

Generates foldlabels from graphs

deep_folding.brainvisa.generate_foldlabels.parse_args(argv)

Parses command-line arguments

Parameters:argv – a list containing command line arguments
Returns:args

deep_folding.brainvisa.generate_skeletons module

Write skeletons from graph files

Typical usage

You can use this program by first entering in the brainvisa environment (here brainvisa 5.0.0 installed with singurity) and launching the script from the terminal: >>> bv bash >>> python generate_skeletons.py

class deep_folding.brainvisa.generate_skeletons.GraphConvert2Skeleton(src_dir, skeleton_dir, side, junction, parallel, path_to_graph, bids, qc_path)

Bases: object

Class to convert all graphs from a folder into skeletons

It contains all information to scan a dataset for graphs and writes skeletons into target directory

compute(number_subjects)

Loops over subjects and converts graphs into skeletons.

generate_one_skeleton(subject: str)

Generates and writes skeleton for one subject.

get_skeleton_filename(subject, graph_file)
deep_folding.brainvisa.generate_skeletons.generate_skeletons(src_dir='/neurospin/dico/data/bv_databases/human/not_labeled/hcp/hcp', skeleton_dir='test/datasets/hcp/skeletons/raw', path_to_graph='t1mri/BL/default_analysis/folds/3.1', side='R', junction='thin', bids=False, parallel=False, number_subjects=-1, qc_path='')

Generates skeletons from graphs

deep_folding.brainvisa.generate_skeletons.parse_args(argv)

Parses command-line arguments

Parameters:argv – a list containing command line arguments
Returns:args

deep_folding.brainvisa.mask_qc module

Quality check of generated masks

deep_folding.brainvisa.mask_qc.check(side='R', sulcus='S.C.', out_voxel_size=(2, 2, 2), nb_subjects=1)
deep_folding.brainvisa.mask_qc.main(argv)

Reads argument line and creates cropped files and pickle file

Parameters:argv – a list containing command line arguments
deep_folding.brainvisa.mask_qc.parse_args(argv)

Function parsing command-line arguments

Parameters:argv – a list containing command line arguments
Returns:dictionary with keys: src_dir, tgt_dir, nb_subjects, list_sulci
Return type:params

deep_folding.brainvisa.pipeline module

deep_folding.brainvisa.pipeline.get_sulci_list(region_name, side, json_path='/neurospin/dico/data/deep_folding/current/sulci_regions_overlap.json')
deep_folding.brainvisa.pipeline.is_directory_empty(path)
deep_folding.brainvisa.pipeline.is_step_to_be_computed(path, log_string, save_behavior='best')
deep_folding.brainvisa.pipeline.parse_args(argv: list) → dict

Function parsing command-line arguments

Parameters:argv – a list containing command line arguments
Returns:a dictionary with all arugments as keys
Return type:params
deep_folding.brainvisa.pipeline.print_info(step, description)

deep_folding.brainvisa.remove_ventricle module

Remove ventricle from a skeleton through the automatic labelled graph by Morphologist. Be careful : the script will not work if there are ‘_’ in subject ids (for database in BIDS format).

class deep_folding.brainvisa.remove_ventricle.RemoveVentricleFromSkeleton(skeleton_dir, output_dir, morpho_dir, path_to_graph, skeleton_filename, output_filename, side, bids, parallel)

Bases: object

Class to remove ventricle from skeleton files through the automatic labelling graph computed by Morphologist. The default automatic labelling session is : deepcnn_session_auto.

It contains all the information to get labelled graphs from skeleton filenames and to write new sketletons without ventricle in the target directory.

compute(number_subjects)

Loops over subjects and remove ventricle from skeletons.

get_labelled_graph(subject: str)

Find the labelled graph in the morphologist database from the skeleton filename.

remove_ventricle_from_one_subject(subject: str)

Removes ventricle and writes new skeleton for one subject.

deep_folding.brainvisa.remove_ventricle.parse_args(argv)

Parses command-line arguments :param argv: a list containing command line arguments

Returns:args
deep_folding.brainvisa.remove_ventricle.remove_ventricle(skeleton_dir='test/datasets/hcp/skeletons/raw', output_dir='test/datasets/hcp/skeletons/raw/without_ventricle', morpho_dir='/neurospin/dico/data/bv_databases/human/not_labeled/hcp/hcp', path_to_graph='t1mri/BL/default_analysis/folds/3.1', skeleton_filename='skeleton_generated', output_filename='skeleton_generated_without_ventricle', side='R', bids=False, parallel=False, number_subjects=-1)

Remove ventricle from a skeleton through the automatic labelled graph by Morphologist

deep_folding.brainvisa.remove_ventricle.remove_ventricle_from_one_skeleton(skeleton_file, labelled_graph_file)

deep_folding.brainvisa.resample_files module

Resample skeletons

The aim of this script is to resample skeletons, foldlabels and distmaps.

You can use this program by first entering in the brainvisa environment (here brainvisa 5.0.0 installed with singurity) and launching the script from the terminal: >>> bv bash >>> python resample_files.py

Alternatively, you can launch the script in the interactive terminal ipython: >>> %run resample_files.py

class deep_folding.brainvisa.resample_files.DistMapResampler(src_dir, resampled_dir, transform_dir, side, out_voxel_size, parallel, src_filename, output_filename)

Bases: deep_folding.brainvisa.resample_files.FileResampler

Resamples all files from source directories

static resample_one_subject(src_file: str, out_voxel_size: float, transform_file: str, resampled_file: str)
class deep_folding.brainvisa.resample_files.FileResampler(src_dir, resampled_dir, transform_dir, side, out_voxel_size, parallel)

Bases: object

Resamples all files from source directories

Parent class from which derive SkeletonResampler, FoldLabelResampler,…

compute(number_subjects=-1)

Loops over nii files

The programm loops over all subjects from the input (source) directory.

Parameters:number_subjects – integer giving the number of subjects to analyze, by default it is set to _ALL_SUBJECTS (-1).
static resample_one_subject(src_file: str, out_voxel_size: float, transform_file: str, resampled_file=None)

Resamples skeleton

This static method is called by resample_one_subject_wrapper from parent class FileResampler

resample_one_subject_wrapper(subject_id)

Resamples one file

Parameters:subject_id – string giving the subject ID
class deep_folding.brainvisa.resample_files.FoldLabelResampler(src_dir, resampled_dir, transform_dir, side, out_voxel_size, parallel, src_filename, output_filename)

Bases: deep_folding.brainvisa.resample_files.FileResampler

Resamples all files from source directories

static resample_one_subject(src_file: str, out_voxel_size: float, transform_file: str, resampled_file=None)
class deep_folding.brainvisa.resample_files.SkeletonResampler(src_dir, resampled_dir, transform_dir, side, out_voxel_size, parallel, src_filename, output_filename)

Bases: deep_folding.brainvisa.resample_files.FileResampler

Resamples all skeletons from source directories

static resample_one_subject(src_file: str, out_voxel_size: float, transform_file: str, resampled_file=None)

Resamples skeleton

This static method is called by resample_one_subject_wrapper from parent class FileResampler

deep_folding.brainvisa.resample_files.parse_args(argv)

Function parsing command-line arguments

Parameters:argv – a list containing command line arguments
Returns:dictionary with keys: src_dir, tgt_dir, nb_subjects, list_sulci
Return type:params
deep_folding.brainvisa.resample_files.resample_files(src_dir='test/datasets/hcp/skeletons/raw', input_type='skeleton', resampled_dir='test/datasets/hcp/skeletons/1mm', transform_dir='test/datasets/hcp/transforms', side='R', out_voxel_size=1.0, parallel=False, number_subjects=-1, src_filename='skeleton_generated_', output_filename='resampled_skeleton_')
deep_folding.brainvisa.resample_files.resample_one_distmap(input_image, resampled_dir, out_voxel_size, transformation)

Resamples one distmap file

Parameters:
  • input_image (either string or aims.Volume) – either path to distmap or distmap aims Volume
  • out_voxel_size (tuple) – Output voxel size (default: None, no resampling)
  • transformation (string or aims.Volume) – either path to transformation file or transformation itself
Returns:

Save transformed or resampled volume
deep_folding.brainvisa.resample_files.resample_one_foldlabel(input_image, out_voxel_size, transformation)

Resamples one foldlabel file

Parameters:
  • input_image (either string or aims.Volume) – either path to foldlabel or foldlabel aims Volume
  • out_voxel_size (tuple) – Output voxel size (default: None, no resampling)
  • transformation (string or aims.Volume) – either path to transformation file or transformation itself
Returns:

aims.Volume

Transformed or resampled volume
Return type:

resampled
deep_folding.brainvisa.resample_files.resample_one_skeleton(input_image, out_voxel_size, transformation)

Resamples one skeleton file

Parameters:
  • input_image (either string or aims.Volume) – either path to skeleton or skeleton aims Volume
  • out_voxel_size (tuple) – Output voxel size (default: None, no resampling)
  • transformation (string or aims.Volume) – either path to transformation file or transformation itself
Returns:

aims.Volume

Transformed or resampled volume
Return type:

resampled

Module contents

deep_folding.brainvisa.exception_handler(func)

This code permits to catch SystemExit with exit code 0

It is a decorator. It is for example used when “–help” is given as argument