monailabel.transform.writer module¶

class monailabel.transform.writer.ClassificationWriter(label='pred', label_names=None)[source]¶: Bases: object

class monailabel.transform.writer.DetectionWriter(json='result', pred_box_key='box', pred_label_key='label', key_write_to_file='result_write_to_file', key_output_format='slicer')[source]¶: Bases: object

class monailabel.transform.writer.PolygonWriter(label='pred', json='result', key_write_to_file='result_write_to_file', key_annotations='annotations', key_label_colors='label_colors', key_output_format='output')[source]¶: Bases: object

class monailabel.transform.writer.Writer(label='pred', json=None, ref_image=None, key_extension='result_extension', key_dtype='result_dtype', key_compress='result_compress', key_write_to_file='result_write_to_file', meta_key_postfix='meta_dict', nibabel=False)[source]¶

Bases: object

is_multichannel_image(image_np)[source]¶

Check if the provided image contains multiple channels

Parameters: image_np (ndarray) – Expected shape (channels, width, height, batch)
Returns: If this is a multi-channel image or not
Return type: bool

monailabel.transform.writer.write_itk(image_np, output_file, affine, dtype, compress)[source]¶

monailabel.transform.writer.write_nifti(data, filename, affine=None, original_affine=None, output_dtype=<class 'numpy.float32'>)[source]¶

monailabel.transform.writer.write_seg_nrrd(image_np, output_file, dtype, affine, labels, color_map=None, index_order='C', space='left-posterior-superior')[source]¶

Write multi-channel seg.nrrd file.

Parameters

image_np (ndarray) – Image as numpy ndarray
output_file (str) – Output file path that the seg.nrrd file should be saved to
dtype (type) – numpy type e.g. float32
affine (ndarray) – Affine matrix
labels (List[str]) – Labels of image segment which will be written to the nrrd header
color_map (Optional[Dict[str, List[float]]]) – Mapping from segment_name(str) to it’s color e.g. {‘heart’: [255/255, 244/255, 209/255]}
index_order (str) – Either ‘C’ or ‘F’ (see nrrd.write() documentation)

Raises

ValueError – In case affine is not provided
ValueError – In case labels are not provided

Return type

None

monailabel.transform.pre module

monailabel.utils package