Description for DataSet class

This class creates imaging dataset objects.

__init__(self, data_path=None, data_type='dcm', mask_names=[], image_only=False, multi_rts_per_pat=False, messy_structure=False, image_names=['image', 'volume', 'img', 'vol']) special

Initialise a dataset object.

Parameters:

  • data_path (str) – Path to DICOM/NRRD root directory (structure inside patient’s folder doesn’t matter).

  • data_type (str) – Original data format, 'dcm' is default, can be 'nrrd'.

  • mask_names (list) – List of names for NRRD files containing binary mask, default is ‘mask’.

  • image_only (bool) – If your dataset has only images, whithout Rtstructures, you need to set it to True, otherwise default value is False.

  • multi_rts_per_pat (bool) – If you have multiple Rtstructures in the patient folder you need to set it to True, otherwise default value is False, to speed up the parsing algorithm.

  • messy_structure (bool) – If data is not organised properly in folders.

  • image_names (list) – List of names for NRRD files containing image file, default is ['image','volume','img','vol'].