MEDscan

MEDscan

class MEDimage.MEDscan.MEDscan(medscan=None)

Bases: object

Organizes all scan data (patientID, imaging data, scan type…).

patientID

Patient ID.

Type:

str

type

Scan type (MRscan, CTscan…).

Type:

str

format

Scan file format. Either ‘npy’ or ‘nifti’.

Type:

str

dicomH

DICOM header.

Type:

pydicom.dataset.FileDataset

data

Instance of MEDscan.data inner class.

Type:

MEDscan.data

class Params

Bases: object

Organizes all processing, filtering and features extraction parameters

class Filter(filter_type: str = '')

Bases: object

Organizes all filtering parameters

class Gabor(sigma: float = 0.0, _lambda: float = 0.0, gamma: float = 0.0, theta: str = '', rot_invariance: bool = False, orthogonal_rot: bool = False, name_save: str = '', padding: str = '')

Bases: object

Organizes the gabor filter parameters

__init__(sigma: float = 0.0, _lambda: float = 0.0, gamma: float = 0.0, theta: str = '', rot_invariance: bool = False, orthogonal_rot: bool = False, name_save: str = '', padding: str = '') → None

Constructor of the Gabor class.

Parameters:

• sigma (float) – Float of the sigma value.

• _lambda (float) – Float of the lambda value.

• gamma (float) – Float of the gamma value.

• theta (str) – String of the theta angle value.

• rot_invariance (bool) – If True the filter will be rotation invariant.

• orthogonal_rot (bool) – If True will compute average response over orthogonal planes.

• name_save (str) – Specific name added to final extraction results file.

• padding (str) – padding mode.

Returns:

None.

init_from_json(params: Dict) → None

Updates class attributes from json file.

Parameters:

params (Dict) – Dictionary of the gabor filter parameters.

Returns:

None.

class Laws(config: List = [], energy_distance: int = 0, energy_image: bool = False, rot_invariance: bool = False, orthogonal_rot: bool = False, name_save: str = '', padding: str = '')

Bases: object

Organizes the laws filter parameters

__init__(config: List = [], energy_distance: int = 0, energy_image: bool = False, rot_invariance: bool = False, orthogonal_rot: bool = False, name_save: str = '', padding: str = '') → None

Constructor of the Laws class.

Parameters:

• config (List) – Configuration of the Laws filter, for ex: [‘E5’, ‘L5’, ‘E5’].

• energy_distance (int) – Chebyshev distance.

• energy_image (bool) – If True will compute the Laws texture energy image.

• rot_invariance (bool) – If True the filter will be rotation invariant.

• orthogonal_rot (bool) – If True will compute average response over orthogonal planes.

• name_save (str) – Specific name added to final extraction results file.

• padding (str) – padding mode.

Returns:

None.

init_from_json(params: Dict) → None

Updates class attributes from json file.

Parameters:

params (Dict) – Dictionary of the laws filter parameters.

Returns:

None.

class Log(ndims: int = 0, sigma: float = 0.0, padding: str = '', orthogonal_rot: bool = False, name_save: str = '')

Bases: object

Organizes the Log filter parameters

__init__(ndims: int = 0, sigma: float = 0.0, padding: str = '', orthogonal_rot: bool = False, name_save: str = '') → None

Constructor of the Log class.

Parameters:

• ndims (int) – Filter dimension.

• sigma (float) – Float of the sigma value.

• padding (str) – padding mode.

• orthogonal_rot (bool) – If True will compute average response over orthogonal planes.

• name_save (str) – Specific name added to final extraction results file.

Returns:

None.

init_from_json(params: Dict) → None

Updates class attributes from json file.

Parameters:

params (Dict) – Dictionary of the Log filter parameters.

Returns:

None.

class Mean(ndims: int = 0, name_save: str = '', padding: str = '', size: int = 0, orthogonal_rot: bool = False)

Bases: object

Organizes the Mean filter parameters

__init__(ndims: int = 0, name_save: str = '', padding: str = '', size: int = 0, orthogonal_rot: bool = False) → None

Constructor of the Mean class.

Parameters:

• ndims (int) – Filter dimension.

• name_save (str) – Specific name added to final extraction results file.

• padding (str) – padding mode.

• size (int) – Filter size.

Returns:

None.

init_from_json(params: Dict) → None

Updates class attributes from json file.

Parameters:

params (Dict) – Dictionary of the Mean filter parameters.

Returns:

None.

class Textural(family: str = '', size: int = 0, discretization: dict = {}, local: bool = False, name_save: str = '')

Bases: object

Organizes the Textural filters parameters

__init__(family: str = '', size: int = 0, discretization: dict = {}, local: bool = False, name_save: str = '') → None

Constructor of the Textural class.

Parameters:

• family (str, optional) – The family of the textural filter.

• size (int, optional) – The filter size.

• discretization (dict, optional) – The discretization parameters.

• local (bool, optional) – If true, the discretization will be computed locally, else globally.

• name_save (str, optional) – Specific name added to final extraction results file.

Returns:

None.

init_from_json(params: Dict) → None

Updates class attributes from json file.

Parameters:

params (Dict) – Dictionary of the wavelet filter parameters.

Returns:

None.

class Wavelet(ndims: int = 0, name_save: str = '', basis_function: str = '', subband: str = '', level: int = 0, rot_invariance: bool = False, padding: str = '')

Bases: object

Organizes the Wavelet filter parameters

__init__(ndims: int = 0, name_save: str = '', basis_function: str = '', subband: str = '', level: int = 0, rot_invariance: bool = False, padding: str = '') → None

Constructor of the Wavelet class.

Parameters:

• ndims (int) – Dimension of the filter.

• name_save (str) – Specific name added to final extraction results file.

• basis_function (str) – Wavelet basis function.

• subband (str) – Wavelet subband.

• level (int) – Decomposition level.

• rot_invariance (bool) – If True the filter will be rotation invariant.

• padding (str) – padding mode.

Returns:

None.

init_from_json(params: Dict) → None

Updates class attributes from json file.

Parameters:

params (Dict) – Dictionary of the wavelet filter parameters.

Returns:

None.

__init__(filter_type: str = '') → None

Constructor of the Filter class.

Parameters:

filter_type (str) – Type of the filter that will be used (Must be ‘mean’, ‘log’, ‘laws’, ‘gabor’ or ‘wavelet’).

Returns:

None.

class Process(**kwargs)

Bases: object

Organizes all processing parameters.

__init__(**kwargs) → None

Constructor of the Process class.

init_from_json(path_to_json: Path | str) → None

Updates class attributes from json file.

Parameters:

path_to_json (Union[Path, str]) – Path to the JSON file with processing parameters.

Returns:

None.

class Radiomics(**kwargs)

Bases: object

Organizes the radiomics extraction parameters

class GLCM(dist_correction: bool | str = False, merge_method: str = 'vol_merge')

Bases: object

Organizes the GLCM features extraction parameters

__init__(dist_correction: bool | str = False, merge_method: str = 'vol_merge') → None

Constructor of the GLCM class

Parameters:

• dist_correction (Union[bool, str]) – norm for distance weighting, must be “manhattan”, “euclidean” or “chebyshev”. If True the norm for distance weighting is gonna be “euclidean”.

• merge_method (str) – merging method which determines how features are calculated. Must be “average”, “slice_merge”, “dir_merge” and “vol_merge”.

Returns:

None.

class GLRLM(dist_correction: bool | str = False, merge_method: str = 'vol_merge')

Bases: object

Organizes the GLRLM features extraction parameters

__init__(dist_correction: bool | str = False, merge_method: str = 'vol_merge') → None

Constructor of the GLRLM class

Parameters:

• dist_correction (Union[bool, str]) – If True the norm for distance weighting is gonna be “euclidean”.

• merge_method (str) – merging method which determines how features are calculated. Must be “average”, “slice_merge”, “dir_merge” and “vol_merge”.

Returns:

None.

class NGTDM(dist_correction: bool | str | None = None)

Bases: object

Organizes the NGTDM features extraction parameters

__init__(dist_correction: bool | str | None = None) → None

Constructor of the NGTDM class

Parameters:

dist_correction (Union[bool, str]) – If True the norm for distance weighting is gonna be “euclidean”.

Returns:

None.

__init__(**kwargs) → None

Constructor of the Radiomics class.

__init__() → None

Organizes all processing, filtering and features extraction

class Radiomics(image: Dict | None = None, params: Dict | None = None)

Bases: object

Organizes all the extracted features.

__init__(image: Dict | None = None, params: Dict | None = None) → None

Constructor of the Radiomics class :param image: Dict of the extracted features. :type image: Dict :param params: Dict of the parameters used in features extraction (roi type, voxels diemension…) :type params: Dict

Returns:

None

to_json() → Dict

Summarizes the class attributes in a Dict :param None:

Returns:

Dictionay of radiomics structure (extracted features and extraction params)

Return type:

Dict

update_params(params: Dict) → None

Updates params attribute from a given Dict :param params: Dict of the parameters used in features extraction (roi type, voxels diemension…) :type params: Dict

Returns:

None

__init__(medscan=None) → None

Constructor of the MEDscan class

Parameters:

medscan (MEDscan) – A MEDscan class instance.

Returns:

None

class data(orientation: str | None = None, patient_position: str | None = None)

Bases: object

Organizes all imaging data (volume and ROI).

volume

Instance of MEDscan.data.volume inner class.

Type:

object

ROI

Instance of MEDscan.data.ROI inner class.

Type:

object

orientation

Imaging data orientation (axial, sagittal or coronal).

Type:

str

patient_position

Patient position specifies the position of the patient relative to the imaging equipment space (HFS, HFP…).

Type:

str

class ROI(indexes: Dict | None = None, roi_names: Dict | None = None)

Bases: object

Organizes all ROI data and information.

indexes

Dict of the ROI indexes for each ROI name.

Type:

Dict

roi_names

Dict of the ROI names.

Type:

Dict

nameSet

Dict of the User-defined name for Structure Set for each ROI name.

Type:

Dict

nameSetInfo

Dict of the names of the structure sets that define the areas of significance. Either ‘StructureSetName’, ‘StructureSetDescription’, ‘SeriesDescription’ or ‘SeriesInstanceUID’.

Type:

Dict

__init__(indexes: Dict | None = None, roi_names: Dict | None = None) → None

Constructor of the ROI class.

Parameters:

• indexes (Dict, optional) – Dict of the ROI indexes for each ROI name.

• roi_names (Dict, optional) – Dict of the ROI names.

Returns:

None.

convert_to_LPS(data: ndarray) → ndarray

Converts the given volume to LPS coordinates system. For more details please refer here : https://www.slicer.org/wiki/Coordinate_systems :param data: Volume data in RAS to convert to to LPS :type data: ndarray

Returns:

n-dimensional of data in LPS.

Return type:

ndarray

get_indexes(key)

get_name_set(key)

get_name_set_info(key)

get_roi_from_path(roi_path: Path | str, id: str)

Extracts all ROI data from the given path for the given patient ID and updates all class attributes with the new extracted data.

Parameters:

• roi_path (Union[Path, str]) – Path where the ROI data is stored.

• id (str) – ID containing patient ID and the modality type, to identify the right file.

Returns:

None.

get_roi_name(key)

update_indexes(key, indexes)

update_name_set(key, name_set)

update_name_set_info(key, nameSetInfo)

update_roi_name(key, roi_name)

__init__(orientation: str | None = None, patient_position: str | None = None) → None

Constructor of the scan class

Parameters:

• orientation (str, optional) – Imaging data orientation (axial, sagittal or coronal).

• patient_position (str, optional) – Patient position specifies the position of the patient relative to the imaging equipment space (HFS, HFP…).

Returns:

None.

display(_slice: int | None = None, roi: str | int = 0) → None

Displays slices from imaging data with the ROI contour in XY-Plane.

Parameters:

• _slice (int, optional) – Index of the slice you want to plot.

• roi (Union[str, int], optional) – ROI name or index. If not specified will use the first ROI.

Returns:

None.

display_process(_slice: int | None = None, roi: str | int = 0) → None

Displays slices from imaging data with the ROI contour in XY-Plane.

Parameters:

• _slice (int, optional) – Index of the slice you want to plot.

• roi (Union[str, int], optional) – ROI name or index. If not specified will use the first ROI.

Returns:

None.

get_indexes_by_roi_name(roi_name: str) → ndarray

Extract ROI data using the ROI name.

Parameters:

roi_name (str) – String of the ROI name (A volume can have multiple ROIs).

Returns:

n-dimensional array of the ROI data.

Return type:

ndarray

get_roi_from_indexes(key: int) → ndarray

Extracts ROI data using the saved indexes (Indexes of non-null values).

Parameters:

key (int) – Key of ROI indexes list (A volume can have multiple ROIs).

Returns:

n-dimensional array of ROI data.

Return type:

ndarray

set_ROI(*args)

set_orientation(orientation)

set_patient_position(patient_position)

set_volume(volume)

class volume(spatialRef: imref3d | None = None, scan_rot: str | None = None, array: ndarray | None = None)

Bases: object

Organizes all volume data and information related to imaging volume.

spatialRef

Imaging data orientation (axial, sagittal or coronal).

Type:

imref3d

scan_rot

Array of the rotation applied to the XYZ points of the ROI.

Type:

ndarray

array

n-dimensional of the imaging data.

Type:

ndarray

__init__(spatialRef: imref3d | None = None, scan_rot: str | None = None, array: ndarray | None = None) → None

Organizes all volume data and information.

Parameters:

• spatialRef (imref3d, optional) – Imaging data orientation (axial, sagittal or coronal).

• scan_rot (ndarray, optional) – Array of the rotation applied to the XYZ points of the ROI.

• array (ndarray, optional) – n-dimensional of the imaging data.

convert_spatialRef()

converts the spatialRef attribute from RAS to LPS coordinates system. <https://www.slicer.org/wiki/Coordinate_systems>.

Parameters:

None.

Returns:

None.

convert_to_LPS()

Convert Imaging data to LPS (Left-Posterior-Superior) coordinates system. <https://www.slicer.org/wiki/Coordinate_systems>.

Returns:

None.

spatialRef_from_nifti(nifti_image_path: Path | str) → None

Computes the imref3d spatialRef using a NIFTI file and updates the spatialRef attribute.

Parameters:

nifti_image_path (str) – String of the NIFTI file path.

Returns:

None.

update_array(array)

update_scan_rot(scan_rot_value)

update_spatialRef(spatialRef_value)

update_transScanToModel(transScanToModel_value)

class volume_process(spatialRef: imref3d | None = None, scan_rot: List | None = None, array: ndarray | None = None, user_string: str = '')

Bases: object

Organizes all volume data and information.

spatialRef

Imaging data orientation (axial, sagittal or coronal).

Type:

imref3d

scan_rot

Array of the rotation applied to the XYZ points of the ROI.

Type:

ndarray

data

n-dimensional of the imaging data.

Type:

ndarray

__init__(spatialRef: imref3d | None = None, scan_rot: List | None = None, array: ndarray | None = None, user_string: str = '') → None

Organizes all volume data and information.

Parameters:

• spatialRef (imref3d, optional) – Imaging data orientation (axial, sagittal or coronal).

• scan_rot (ndarray, optional) – Array of the rotation applied to the XYZ points of the ROI.

• array (ndarray, optional) – n-dimensional of the imaging data.

• user_string (str, optional) – string explaining the processed data in the class.

Returns:

None.

load(file_name: str, loading_path: Path | str, update: bool = True) → None | ndarray

Saves the processed data locally.

Parameters:

• file_name (str) – Name file of the processed data to load.

• loading_path (Union[Path, str]) – Path to the processed data to load.

• update (bool, optional) – If True, updates the class attrtibutes with loaded data.

Returns:

None.

save(name_save: str, path_save: Path | str) → None

Saves the processed data locally.

Parameters:

• name_save (str) – Saving name of the processed data.

• path_save (Union[Path, str]) – Path to where save the processed data.

Returns:

None.

update_processed_data(array: ndarray, user_string: str = '') → None

init_from_nifti(nifti_image_path: Path) → None

Initializes the MEDscan class using a NIfTI file.

Parameters:

nifti_image_path (Path) – NIfTI file path.

Returns:

None.

init_ntf_calculation(vol_obj: image_volume_obj) → None

Initializes all the computation parameters for non-texture features as well as the results dict.

Parameters:

vol_obj (image_volume_obj) – Imaging volume.

Returns:

None.

init_params(im_param_scan: Dict) → None

Initializes the Params class from a dictionary.

Parameters:

im_param_scan (Dict) – Dictionary of different processing, extraction and filtering params.

Returns:

None.

init_tf_calculation(algo: int, gl: int, scale: int) → None

Initializes all the computation parameters for the texture-features as well as the results dict.

Parameters:

• algo (int) – Discretisation algorithms index.

• gl (int) – gray-level index.

• scale (int) – scale-text index.

Returns:

None.

save_radiomics(scan_file_name: List, path_save: Path, roi_type: str, roi_type_label: str, patient_num: int | None = None) → None

Saves extracted radiomics features in a JSON file.

Parameters:

• scan_file_name (List) – List of scan files.

• path_save (Path) – Saving path.

• roi_type (str) – Type of the ROI.

• roi_type_label (str) – Label of the ROI type.

• patient_num (int) – Index of scan.

Returns:

None.

update_radiomics(int_vol_hist_features: Dict = {}, morph_features: Dict = {}, loc_int_features: Dict = {}, stats_features: Dict = {}, int_hist_features: Dict = {}, glcm_features: Dict = {}, glrlm_features: Dict = {}, glszm_features: Dict = {}, gldzm_features: Dict = {}, ngtdm_features: Dict = {}, ngldm_features: Dict = {}) → None

Updates the results attribute with the extracted features.

Parameters:

• int_vol_hist_features (Dict, optional) – Dictionary of the intensity volume histogram features.

• morph_features (Dict, optional) – Dictionary of the morphological features.

• loc_int_features (Dict, optional) – Dictionary of the intensity local intensity features.

• stats_features (Dict, optional) – Dictionary of the statistical features.

• int_hist_features (Dict, optional) – Dictionary of the intensity histogram features.

• glcm_features (Dict, optional) – Dictionary of the GLCM features.

• glrlm_features (Dict, optional) – Dictionary of the GLRLM features.

• glszm_features (Dict, optional) – Dictionary of the GLSZM features.

• gldzm_features (Dict, optional) – Dictionary of the GLDZM features.

• ngtdm_features (Dict, optional) – Dictionary of the NGTDM features.

• ngldm_features (Dict, optional) – Dictionary of the NGLDM features.

Returns:

None.