dcnum.segm.segm_torch.torch_postproc
====================================

.. py:module:: dcnum.segm.segm_torch.torch_postproc


Attributes
----------

.. autoapisummary::

   dcnum.segm.segm_torch.torch_postproc.ndi


Functions
---------

.. autoapisummary::

   dcnum.segm.segm_torch.torch_postproc.postprocess_masks


Module Contents
---------------

.. py:data:: ndi

.. py:function:: postprocess_masks(masks, original_image_shape: tuple[int, int])

   Postprocess mask images from ML segmenters

   The transformation includes:

   - Revert the cropping and padding operations done in
     :func:`.preprocess_images` by padding with zeros and cropping.
   - If the original image shape is larger than the mask image shape,
     also clear borders in an intermediate step
     (maks postprocessing using :func:`Segmenter.process_labels`).

   :param masks: Mask data in shape (batch_size, 1, imagex_size, imagey_size)
                 or (batch_size, imagex_size, imagey_size).
   :type masks: 3d or 4d ndarray
   :param original_image_shape: The required output mask shape for one event. This required for
                                doing the inverse of what is done in :func:`.preprocess_images`.
   :type original_image_shape: tuple of (int, int)

   :returns: **labels_proc** -- An integer array with the same dimensions as the original image
             data passed to :func:`.preprocess_images`. The shape of this array
             is (batch_size, original_image_shape[0], original_image_shape[1]).
   :rtype: np.ndarray

   .. rubric:: Notes

   This method is only called by the overarching logic when the
   preprocessing/model output produces images of different shape.
   It causes an obvious overhead that we want to avoid.