searchFirst/processing_methods.py at main · MaksHess/searchFirst · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
import logging
from pathlib import Path
import napari
import numpy as np
from skimage import filters
from skimage.measure import label, regionprops
from skimage.feature import match_template
from skimage.morphology import local_maxima, remove_small_objects
from scipy.ndimage import zoom
import imageio


def find_objects_by_template_matching(stitched_ds, object_threshold, template_path, downsampling, well,
                                      n_objects_per_site):
    if template_path is None:
        template_path = Path(r'C:\Users\CVUser\Documents\Python\searchFirst\templates\template_ZE_9x.tif')
    logging.info(f"loading template from {template_path}...")
    template = imageio.imread(template_path)
    template_ds = zoom(template, downsampling)

    match = match_template(stitched_ds, template_ds, pad_input=True, mode='constant', constant_values=100)
    match_thresholded = np.where(match > object_threshold, match, 0)
    if np.sum(match_thresholded) == 0:
        logging.warning(
            f"no matches found in {well}! Try lowering the `object_threshold` if you expected to find matches in this well.")
        return np.zeros_like(stitched_ds), np.zeros_like(stitched_ds)
    maxima = local_maxima(match_thresholded)
    n_objects = np.sum(maxima)
    logging.info(f'{n_objects} objects found...')
    score = match[np.where(maxima)]
    n_actual = n_objects_per_site
    if n_objects < n_objects_per_site:
        logging.warning(f"only {n_objects} objects found instead of {n_objects_per_site}")
        n_actual = n_objects
    nth_largest_score = -np.partition(-score, n_actual - 1)[n_actual - 1]
    weighted_maxima = np.where(maxima, match, 0)
    selected_objects = np.where(weighted_maxima >= nth_largest_score, weighted_maxima, 0)
    unselected_objects = np.where(np.logical_and(weighted_maxima > 0, weighted_maxima < nth_largest_score),
                                  weighted_maxima, 0)
    return selected_objects, unselected_objects


def find_objects_by_multiple_template_matching(stitched_ds, object_threshold,
                                               template_path, downsampling,
                                               well, n_objects_per_site):
    if template_path is None:
        template_path = Path((r'C:\Users\CVUser\Documents\Python\searchFirst'
                              '\templates\template_ZE_9x.tif'))
    logging.info(f"loading template from {template_path}...")

    # get list of template files
    template_path = Path(template_path)
    template_files = template_path.glob('*.tif')

    # initialize arrays for object matches from all templates combined
    all_selected = np.empty(np.shape(stitched_ds))
    all_unselected = np.empty(np.shape(stitched_ds))

    # iterate over templates
    for fyle in template_files:
        template = imageio.imread(fyle)
        template_ds = zoom(template, downsampling)

        match = match_template(stitched_ds, template_ds, pad_input=True,
                               mode='constant', constant_values=100)
        match_thresholded = np.where(match > object_threshold, match, 0)
        if np.sum(match_thresholded) == 0:
            logging.warning(
                f"no matches found in {well} for template {fyle}! '"
                f"'Try lowering the `object_threshold` if you expected to'"
                f"' find matches in this well.")
            continue
        maxima = local_maxima(match_thresholded)
        n_objects = np.sum(maxima)
        logging.info(f'{n_objects} objects found for template {fyle}...')
        score = match[np.where(maxima)]
        n_actual = n_objects_per_site
        if n_objects < n_objects_per_site:
            logging.warning(f"only {n_objects} objects found instead of '"
                            f"'{n_objects_per_site}")
            n_actual = n_objects

        nth_largest_score = -np.partition(-score, n_actual - 1)[n_actual - 1]
        weighted_maxima = np.where(maxima, match, 0)
        selected_objects = np.where(weighted_maxima >= nth_largest_score,
                                    weighted_maxima, 0)
        unselected_objects = np.where(
            np.logical_and(weighted_maxima > 0,
                           weighted_maxima < nth_largest_score),
            weighted_maxima, 0)

        all_selected += selected_objects
        all_unselected += unselected_objects

    return all_selected, all_unselected


def find_objects_by_threshold(stitched_ds, sigma, minimum_object_size):
    # Normalize stitched_ds
    img = stitched_ds / np.amax(stitched_ds)
    # initialize canvas of zeroes
    selected_objects = np.zeros(img.shape)

    gaussian = filters.gaussian(img, sigma=sigma)

    threshold_gaussian = filters.threshold_otsu(gaussian)

    binary_gaussian = gaussian >= threshold_gaussian

    masked = remove_small_objects(binary_gaussian, minimum_object_size)

    labeled_blobs = label(masked)

    props = regionprops(labeled_blobs)

    for props in props:
        a = props.centroid
        selected_objects[int(a[0]), int(a[1])] = 1

    return selected_objects, np.zeros_like(selected_objects)


def find_objects_by_manual_annotation(stitched_ds):
    viewer = napari.Viewer()
    viewer.add_image(stitched_ds)
    # rescale stitched image
    low, high = np.quantile(stitched_ds, [0.0001, 0.9999])
    viewer.layers['stitched_ds'].contrast_limits = [low, high]
    viewer.add_points(None)
    viewer.layers['Points'].mode = 'add'
    viewer.show(block=True)

    # after the viewer is closed, the following will be executed:
    coords = viewer.layers['Points'].data
    n_objects = len(coords)
    if n_objects == 0:
        logging.warning('no coordinates were annotated...')
    else:
        logging.info(f'{n_objects} coordinates were annotated...')
    selected_objects = np.empty(np.shape(stitched_ds))
    selected_objects[coords[:, 0].astype('int'),
                     coords[:, 1].astype('int')] = 1
    unselected_objects = np.empty(np.shape(stitched_ds))

    return selected_objects, unselected_objects


def find_objects_by_semiautomatic_annotation(stitched_ds, sigma, minimum_object_size):
    objects, non_objects = find_objects_by_threshold(stitched_ds,
                                                     sigma,
                                                     minimum_object_size,
                                                     )
    points = np.stack(np.where(objects)).T

    viewer = napari.Viewer()
    viewer.add_image(stitched_ds)
    # rescale stitched image
    low, high = np.quantile(stitched_ds, [0.0001, 0.9999])
    viewer.layers['stitched_ds'].contrast_limits = [low, high]
    viewer.add_points(points, symbol='ring', size=100)
    viewer.layers['Points'].mode = 'add'
    viewer.show(block=True)

    # after the viewer is closed, the following will be executed:
    coords = viewer.layers['Points'].data
    n_objects = len(coords)
    if n_objects == 0:
        logging.warning('no coordinates were annotated...')
    else:
        logging.info(f'{n_objects} coordinates were annotated...')
    selected_objects = np.empty(np.shape(stitched_ds))
    selected_objects[coords[:, 0].astype('int'),
                     coords[:, 1].astype('int')] = 1
    unselected_objects = np.empty(np.shape(stitched_ds))

    return selected_objects, unselected_objects