We noticed that the error occured when the grandchallenge load the prediced results. We think it may because of the generation process of prediction results, we provide a simple example, which we have testes to be able to sucessfully submit.
def out(segmentation, output_path, image_path):
"""
save your predictions
:param segmentation:Your prediction , the data type is "numpy.ndarray".
:param output_path:The save path of prediction results.
:param image_path:The path of the image corresponding to the prediction result.
:return:
"""
dicom = sitk.ReadImage(image_path)
output = sitk.GetImageFromArray(segmentation)
output.SetOrigin(dicom.GetOrigin())
output.SetSpacing(dicom.GetSpacing())
output.SetDirection(dicom.GetDirection())
sitk.WriteImage(output, output_path)