occlusions
==========

.. py:module:: occlusions


Functions
---------

.. autoapisummary::

   occlusions.__split_groups
   occlusions.spectral_occlusion
   occlusions.context_occlusion


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

.. py:function:: __split_groups(neg_mask)

.. py:function:: spectral_occlusion(mask, data, noise=0.02, device = 'cpu')

   Linear interpolated occlusion for spectral data, with optional added noise.

   @param mask: boolean valued NDArray
   @param data: data to be occluded
   @param noise: parameter for optional gaussian noise.
       Set to 0.0 if you want simple linear interpolation

   @return torch.Tensor


.. py:function:: context_occlusion(mask, data, context, noise=0.5)

   Context based occlusion with optional added noise.

   @param mask: boolean valued NDArray
   @param data: data to be occluded
   @param context: data to be used as occlusion e.g. CT scan of a healthy patient or a road
   @param noise: parameter for optional gaussian noise.
       Set to 0.0 for no noise

   @return torch.Tensor