"""
BiaPy post-processing package.
This package provides post-processing utilities and functions for refining model
predictions, such as filtering, morphological operations, and other enhancements
for 2D and 3D biomedical image data.
"""
from biapy.data.post_processing.post_processing import apply_median_filtering
[docs]
def apply_post_processing(cfg, data):
"""
Create training and validation generators.
Parameters
----------
cfg : YACS CN object
Configuration.
data : 4D/5D Numpy array
Data to apply post_proccessing. E.g. ``(num_of_images, y, x, channels)`` for 2D and
``(num_of_images, z, y, x, channels)`` for 3D.
Returns
-------
data : 4D/5D Numpy array
Data to apply post_proccessing. E.g. ``(num_of_images, y, x, channels)`` for 2D and
``(num_of_images, z, y, x, channels)`` for 3D.
"""
print("Applying post-processing . . .")
for f, val in zip(
cfg.TEST.POST_PROCESSING.MEDIAN_FILTER_AXIS,
cfg.TEST.POST_PROCESSING.MEDIAN_FILTER_SIZE,
):
data = apply_median_filtering(data, axes=f, mf_size=val)
return data