rs2mask

Dataset

class segmentation_rt.rs2mask.rs2mask.Dataset(path, export_path, structures, force=True)

Bases: object

From dicom to dataset class. Convert CT and RTSTRUCT into nii, readable by deep learning frameworks.

All subfolders representing subject must contain the CT and the RS associated.

Example:
>>> from segmentation_rt.rs2mask import Dataset
>>> structures = ['Heart', 'Breast L', 'Breast R']
>>> dataset = Dataset('data/dicom_dataset', 'data/nii_dataset', structures)
>>> dataset.make()
Parameters

  • path (string) – Root directory.

  • export_path (string) – Export path.

  • structures (list[string]) – List of desired structure(s).

  • force (bool) – Force export even if one structure is missing.

find_structures(index)

List missing and not missing structures in a RTSTRUCT.

Parameters

index (int) – index of the patient.

Returns

List missing and not missing structures.

Return type

(list[str],list[str])

get_rs()

List RTSTRUCT for each patient.

Return type

list[str]

make()

Create structures and convert the CT in nii format for each subject.