input.py

# Copyright CNRS/Inria/UNS
# Contributor(s): Eric Debreuve (since 2019), Morgane Nadal (2020)
#
# eric.debreuve@cnrs.fr
#
# This software is governed by the CeCILL  license under French law and
# abiding by the rules of distribution of free software.  You can  use,
# modify and/ or redistribute the software under the terms of the CeCILL
# license as circulated by CEA, CNRS and INRIA at the following URL
# "http://www.cecill.info".
#
# As a counterpart to the access to the source code and  rights to copy,
# modify and redistribute granted by the license, users are provided only
# with a limited warranty  and the software's author,  the holder of the
# economic rights,  and the successive licensors  have only  limited
# liability.
#
# In this respect, the user's attention is drawn to the risks associated
# with loading,  using,  modifying and/or developing or reproducing the
# software by the user in light of its specific status of free software,
# that may mean  that it is complicated to manipulate,  and  that  also
# therefore means  that it is reserved for developers  and  experienced
# professionals having in-depth computer knowledge. Users are therefore
# encouraged to load and test the software's suitability as regards their
# requirements in conditions enabling the security of their systems and/or
# data to be ensured and,  more generally, to use and operate it in the
# same conditions as regards security.
#
# The fact that you are presently reading this means that you have had
# knowledge of the CeCILL license and that you accept its terms.

from brick.general.type import array_t
import numpy as np_
import sys as sy_
from PIL import Image
from PIL.ExifTags import TAGS


def ImageVerification(image: array_t, channel: str) -> array_t:
    #
    # The image must not be constant
    if image.max() == image.min():
        raise ValueError('The input image should not be constant.')

    # Verification of the dimension of the image and its coherence with the parameters (channel)
    elif image.ndim == 3:
        print('Your image has only one color channel.')
        if channel is not None:
            raise ValueError('The image has only 3 dimensions. However, a value for the "channel" parameter is '
                             'specified. Give the channel the value None')

    elif image.ndim == 4:
        if channel == 'R' or channel == 'G' or channel == 'B':
            # not changed into if --channel in 'RGB'-- because if channel='RG' => True.
            print('The image has multiple color channels. The channel', channel,
                  'is specified in the parameters.')
            image = image[:, :, :, 'RGB'.find(channel)]

            # The obtained image must not be constant
            if image.max() == image.min():
                raise ValueError('The input image should not be constant.')

            return image

        else:
            raise ValueError('The image has multiple color channels. Error in the value of the parameter channel.')

    elif image.ndim != 4 and image.ndim != 3:
        raise ValueError(f'The image dimensions are not correct: {image.ndim}, instead of 3 or 4.')


def IntensityNormalizedImage(image: array_t) -> array_t:
    #
    print('Relative Intensity Normalization between 0 and 1. Need to reevaluate the parameters !!!')

    value_max = image.max()
    value_min = image.min()

    result = (image.astype(np_.float32) - value_min) / float(value_max - value_min)

    return result

    # print(
    #     "This normalization does not bring anything; left as is for now to avoid the need for changing prms"
    # )
    # nonextreme_values = image[np_.logical_and(image > 0.0, image < image.max())]
    #
    # if nonextreme_values.size > 0:
    #     nonextreme_avg = nonextreme_values.mean()
    #     result = image.astype(np_.float32) / nonextreme_avg
    # else:
    #     result = image.astype(np_.float32)
    #
    # return result


def VoxelDimensionInMicrons(data_path: str) -> array_t:
    # TODO Verify if metadata are in the same format for all the images - if not, add a raise error ?

    with Image.open(data_path) as img:  # Find the voxels dimensions in micron in the metadata.
        meta_dict = {TAGS.get(key, 'missing'): img.tag[key] for key in
                     img.tag}  # Use the exif tags into the image metadata

    # Decode the tags text
    metadata = meta_dict['missing'].decode('utf8')  # Decode the tags text
    metadata = metadata.replace('\x00', '')

    voxel_size = ['', '', '']  # Initialize the list of voxel size in str
    for idaxe, axe in enumerate('XYZ'):
        idvox = metadata.find('dblVoxel' + axe)  # Get the index of the voxel size for the axe X, Y, Z
        idvox += 15
        while metadata[idvox] != '\n':
            voxel_size[idaxe] += metadata[idvox]
            idvox += 1
    voxel_size = np_.asarray(list(map(float, voxel_size)))
    voxel_size_microns = 1.0e6 * voxel_size
    print('Voxel dimension in the image is [X Y Z] =', voxel_size_microns, 'in microns. WARNING this method highly '
                                                                           'depends on the format of the metadata.')

    return voxel_size_microns


def ToPixels(micron: float, voxel_size_microns: array_t) -> int:
    # Conversion of pixels into microns. Used in morphomath (disk structuring element mp_.disk(n_pixel)).
    # Assumes that the axes X and Y have the same ratio pixel:microns.
    return round(micron/voxel_size_microns[0])


def ToMicrons(pixel: int, voxel_size_microns: array_t) -> float:
    # May not be used. if not used, delete.
    # Conversion of microns into pixels.
    # Assumes that the axes X and Y have the same ratio pixel:microns.
    return float(pixel * voxel_size_microns[0])


def DijkstraCosts(image: array_t, som_map: array_t, ext_map: array_t) -> array_t:
    #
    # TODO: Ideally, the extension part should be dilated
    # but in ext-ext connections, there must not be dilation around the current or the other exts
    # (current ext plays the role of a soma in soma-ext step)
    #
    dijkstra_costs = 1.0 / (image + 1.0)
    dijkstra_costs[np_.logical_or(som_map > 0, ext_map > 0)] = np_.inf

    return dijkstra_costs