-
Notifications
You must be signed in to change notification settings - Fork 71
/
extract_feature_heatmap.py
375 lines (288 loc) · 16.7 KB
/
extract_feature_heatmap.py
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
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
import csv
import glob
import os
import random
import cv2
import numpy as np
import scipy.stats.stats as st
from skimage.measure import label
from skimage.measure import regionprops
from camelyon16 import utils as utils
from camelyon16.ops.wsi_ops import WSIOps
FILTER_DIM = 2
N_FEATURES = 31
MAX, MEAN, VARIANCE, SKEWNESS, KURTOSIS = 0, 1, 2, 3, 4
def get_region_props(heatmap_threshold_2d, heatmap_prob_2d):
labeled_img = label(heatmap_threshold_2d)
return regionprops(labeled_img, intensity_image=heatmap_prob_2d)
def draw_bbox(heatmap_threshold, region_props, threshold_label='t90'):
n_regions = len(region_props)
print('No of regions(%s): %d' % (threshold_label, n_regions))
for index in range(n_regions):
# print('\n\nDisplaying region: %d' % index)
region = region_props[index]
# print('area: ', region['area'])
# print('bbox: ', region['bbox'])
# print('centroid: ', region['centroid'])
# print('convex_area: ', region['convex_area'])
# print('eccentricity: ', region['eccentricity'])
# print('extent: ', region['extent'])
# print('major_axis_length: ', region['major_axis_length'])
# print('minor_axis_length: ', region['minor_axis_length'])
# print('orientation: ', region['orientation'])
# print('perimeter: ', region['perimeter'])
# print('solidity: ', region['solidity'])
cv2.rectangle(heatmap_threshold, (region['bbox'][1], region['bbox'][0]),
(region['bbox'][3], region['bbox'][2]), color=(0, 255, 0),
thickness=1)
cv2.ellipse(heatmap_threshold, (int(region['centroid'][1]), int(region['centroid'][0])),
(int(region['major_axis_length'] / 2), int(region['minor_axis_length'] / 2)),
region['orientation'] * 90, 0, 360, color=(0, 0, 255),
thickness=2)
cv2.imshow('bbox_%s' % threshold_label, heatmap_threshold)
def get_largest_tumor_index(region_props):
largest_tumor_index = -1
largest_tumor_area = -1
n_regions = len(region_props)
for index in range(n_regions):
if region_props[index]['area'] > largest_tumor_area:
largest_tumor_area = region_props[index]['area']
largest_tumor_index = index
return largest_tumor_index
def get_longest_axis_in_largest_tumor_region(region_props, largest_tumor_region_index):
largest_tumor_region = region_props[largest_tumor_region_index]
return max(largest_tumor_region['major_axis_length'], largest_tumor_region['minor_axis_length'])
def get_tumor_region_to_tissue_ratio(region_props, image_open):
tissue_area = cv2.countNonZero(image_open)
tumor_area = 0
n_regions = len(region_props)
for index in range(n_regions):
tumor_area += region_props[index]['area']
return float(tumor_area) / tissue_area
def get_tumor_region_to_bbox_ratio(region_props):
# for all regions or largest region
print()
def get_feature(region_props, n_region, feature_name):
feature = [0] * 5
if n_region > 0:
feature_values = [region[feature_name] for region in region_props]
feature[MAX] = utils.format_2f(np.max(feature_values))
feature[MEAN] = utils.format_2f(np.mean(feature_values))
feature[VARIANCE] = utils.format_2f(np.var(feature_values))
feature[SKEWNESS] = utils.format_2f(st.skew(np.array(feature_values)))
feature[KURTOSIS] = utils.format_2f(st.kurtosis(np.array(feature_values)))
return feature
def get_average_prediction_across_tumor_regions(region_props):
# close 255
region_mean_intensity = [region.mean_intensity for region in region_props]
return np.mean(region_mean_intensity)
def extract_features(heatmap_prob, image_open):
"""
Feature list:
-> (01) given t = 0.90, total number of tumor regions
-> (02) given t = 0.90, percentage of tumor region over the whole tissue region
-> (03) given t = 0.50, the area of largest tumor region
-> (04) given t = 0.50, the longest axis in the largest tumor region
-> (05) given t = 0.90, total number pixels with probability greater than 0.90
-> (06) given t = 0.90, average prediction across tumor region
-> (07-11) given t = 0.90, max, mean, variance, skewness, and kurtosis of 'area'
-> (12-16) given t = 0.90, max, mean, variance, skewness, and kurtosis of 'perimeter'
-> (17-21) given t = 0.90, max, mean, variance, skewness, and kurtosis of 'compactness(eccentricity[?])'
-> (22-26) given t = 0.50, max, mean, variance, skewness, and kurtosis of 'rectangularity(extent)'
-> (27-31) given t = 0.90, max, mean, variance, skewness, and kurtosis of 'solidity'
:param heatmap_prob:
:param image_open:
:return:
"""
heatmap_threshold_t90 = np.array(heatmap_prob)
heatmap_threshold_t50 = np.array(heatmap_prob)
heatmap_threshold_t90[heatmap_threshold_t90 < int(0.90 * 255)] = 0
heatmap_threshold_t90[heatmap_threshold_t90 >= int(0.90 * 255)] = 255
heatmap_threshold_t50[heatmap_threshold_t50 <= int(0.50 * 255)] = 0
heatmap_threshold_t50[heatmap_threshold_t50 > int(0.50 * 255)] = 255
heatmap_threshold_t90_2d = np.reshape(heatmap_threshold_t90[:, :, :1],
(heatmap_threshold_t90.shape[0], heatmap_threshold_t90.shape[1]))
heatmap_threshold_t50_2d = np.reshape(heatmap_threshold_t50[:, :, :1],
(heatmap_threshold_t50.shape[0], heatmap_threshold_t50.shape[1]))
heatmap_prob_2d = np.reshape(heatmap_prob[:, :, :1],
(heatmap_prob.shape[0], heatmap_prob.shape[1]))
region_props_t90 = get_region_props(np.array(heatmap_threshold_t90_2d), heatmap_prob_2d)
region_props_t50 = get_region_props(np.array(heatmap_threshold_t50_2d), heatmap_prob_2d)
features = []
f_count_tumor_region = len(region_props_t90)
if f_count_tumor_region == 0:
return [0.00] * N_FEATURES
features.append(utils.format_2f(f_count_tumor_region))
f_percentage_tumor_over_tissue_region = get_tumor_region_to_tissue_ratio(region_props_t90, image_open)
features.append(utils.format_2f(f_percentage_tumor_over_tissue_region))
largest_tumor_region_index_t90 = get_largest_tumor_index(region_props_t90)
largest_tumor_region_index_t50 = get_largest_tumor_index(region_props_t50)
f_area_largest_tumor_region_t50 = region_props_t50[largest_tumor_region_index_t50].area
features.append(utils.format_2f(f_area_largest_tumor_region_t50))
f_longest_axis_largest_tumor_region_t50 = get_longest_axis_in_largest_tumor_region(region_props_t50,
largest_tumor_region_index_t50)
features.append(utils.format_2f(f_longest_axis_largest_tumor_region_t50))
f_pixels_count_prob_gt_90 = cv2.countNonZero(heatmap_threshold_t90_2d)
features.append(utils.format_2f(f_pixels_count_prob_gt_90))
f_avg_prediction_across_tumor_regions = get_average_prediction_across_tumor_regions(region_props_t90)
features.append(utils.format_2f(f_avg_prediction_across_tumor_regions))
f_area = get_feature(region_props_t90, f_count_tumor_region, 'area')
features += f_area
f_perimeter = get_feature(region_props_t90, f_count_tumor_region, 'perimeter')
features += f_perimeter
f_eccentricity = get_feature(region_props_t90, f_count_tumor_region, 'eccentricity')
features += f_eccentricity
f_extent_t50 = get_feature(region_props_t50, len(region_props_t50), 'extent')
features += f_extent_t50
f_solidity = get_feature(region_props_t90, f_count_tumor_region, 'solidity')
features += f_solidity
# f_longest_axis_largest_tumor_region_t90 = get_longest_axis_in_largest_tumor_region(region_props_t90,
# largest_tumor_region_index_t90)
# f_area_larget_tumor_region_t90 = region_props_t90[largest_tumor_region_index_t90].area
# cv2.imshow('heatmap_threshold_t90', heatmap_threshold_t90)
# cv2.imshow('heatmap_threshold_t50', heatmap_threshold_t50)
# draw_bbox(np.array(heatmap_threshold_t90), region_props_t90, threshold_label='t90')
# draw_bbox(np.array(heatmap_threshold_t50), region_props_t50, threshold_label='t50')
# key = cv2.waitKey(0) & 0xFF
# if key == 27: # escape
# exit(0)
return features
def extract_features_test(heatmap_prob_name_postfix_first_model, heatmap_prob_name_postfix_second_model, f_test):
print('************************** extract_features_test() ***************************')
print('heatmap_prob_name_postfix_first_model: %s' % heatmap_prob_name_postfix_first_model)
print('heatmap_prob_name_postfix_second_model: %s' % heatmap_prob_name_postfix_second_model)
print('f_test: %s' % f_test)
test_wsi_paths = glob.glob(os.path.join(utils.TEST_WSI_PATH, '*.tif'))
test_wsi_paths.sort()
features_file_test = open(f_test, 'w')
wr_test = csv.writer(features_file_test, quoting=csv.QUOTE_NONNUMERIC)
wr_test.writerow(utils.heatmap_feature_names[:len(utils.heatmap_feature_names) - 1])
for wsi_path in test_wsi_paths:
wsi_name = utils.get_filename_from_path(wsi_path)
print('extracting features for: %s' % wsi_name)
heatmap_prob_path = glob.glob(
os.path.join(utils.HEAT_MAP_DIR, '*%s*%s' % (wsi_name, heatmap_prob_name_postfix_first_model)))
# print(heatmap_prob_path)
image_open = wsi_ops.get_image_open(wsi_path)
heatmap_prob = cv2.imread(heatmap_prob_path[0])
if heatmap_prob_name_postfix_second_model is not None:
heatmap_prob_path_second_model = glob.glob(
os.path.join(utils.HEAT_MAP_DIR, '*%s*%s' % (wsi_name, heatmap_prob_name_postfix_second_model)))
heatmap_prob_second_model = cv2.imread(heatmap_prob_path_second_model[0])
for row in range(heatmap_prob.shape[0]):
for col in range(heatmap_prob.shape[1]):
if heatmap_prob[row, col, 0] >= 0.90 * 255 and heatmap_prob_second_model[row, col, 0] < 0.50 * 255:
heatmap_prob[row, col, :] = heatmap_prob_second_model[row, col, :]
features = extract_features(heatmap_prob, image_open)
print(features)
wr_test.writerow(features)
def extract_features_train_all(heatmap_prob_name_postfix_first_model, heatmap_prob_name_postfix_second_model, f_train):
print('********************** extract_features_train_all() *************************')
print('heatmap_prob_name_postfix_first_model: %s' % heatmap_prob_name_postfix_first_model)
print('heatmap_prob_name_postfix_second_model: %s' % heatmap_prob_name_postfix_second_model)
print('f_train: %s' % f_train)
tumor_wsi_paths = glob.glob(os.path.join(utils.TUMOR_WSI_PATH, '*.tif'))
tumor_wsi_paths.sort()
normal_wsi_paths = glob.glob(os.path.join(utils.NORMAL_WSI_PATH, '*.tif'))
normal_wsi_paths.sort()
wsi_paths = tumor_wsi_paths + normal_wsi_paths
features_file_train_all = open(f_train, 'w')
wr_train = csv.writer(features_file_train_all, quoting=csv.QUOTE_NONNUMERIC)
wr_train.writerow(utils.heatmap_feature_names)
for wsi_path in wsi_paths:
wsi_name = utils.get_filename_from_path(wsi_path)
# print('extracting features for: %s' % wsi_name)
heatmap_prob_path = glob.glob(
os.path.join(utils.HEAT_MAP_DIR, '*%s*%s' % (wsi_name, heatmap_prob_name_postfix_first_model)))
# print(heatmap_prob_path)
image_open = wsi_ops.get_image_open(wsi_path)
heatmap_prob = cv2.imread(heatmap_prob_path[0])
if heatmap_prob_name_postfix_second_model is not None:
heatmap_prob_path_second_model = glob.glob(
os.path.join(utils.HEAT_MAP_DIR, '*%s*%s' % (wsi_name, heatmap_prob_name_postfix_second_model)))
heatmap_prob_second_model = cv2.imread(heatmap_prob_path_second_model[0])
for row in range(heatmap_prob.shape[0]):
for col in range(heatmap_prob.shape[1]):
if heatmap_prob[row, col, 0] >= 0.90 * 255 and heatmap_prob_second_model[row, col, 0] < 0.50 * 255:
heatmap_prob[row, col, :] = heatmap_prob_second_model[row, col, :]
features = extract_features(heatmap_prob, image_open)
if 'umor' in wsi_name:
features += [1]
else:
features += [0]
print(features)
wr_train.writerow(features)
def extract_features_train_validation(heatmap_prob_name_postfix_first_model, heatmap_prob_name_postfix_second_model,
f_train, f_validation):
print('********************** extract_features_train_validation() ********************************')
print('heatmap_prob_name_postfix_first_model: %s' % heatmap_prob_name_postfix_first_model)
print('heatmap_prob_name_postfix_second_model: %s' % heatmap_prob_name_postfix_second_model)
print('f_train: %s' % f_train)
print('f_validation: %s' % f_validation)
tumor_wsi_paths = glob.glob(os.path.join(utils.TUMOR_WSI_PATH, '*.tif'))
tumor_wsi_paths.sort()
normal_wsi_paths = glob.glob(os.path.join(utils.NORMAL_WSI_PATH, '*.tif'))
normal_wsi_paths.sort()
tumor_shuffled_index = list(range(len(tumor_wsi_paths)))
random.seed(12345)
random.shuffle(tumor_shuffled_index)
normal_shuffled_index = list(range(len(tumor_wsi_paths), len(tumor_wsi_paths) + len(normal_wsi_paths)))
random.seed(12345)
random.shuffle(normal_shuffled_index)
tumor_shuffled_index = tumor_shuffled_index[:20]
normal_shuffled_index = normal_shuffled_index[:30]
validation_index = tumor_shuffled_index + normal_shuffled_index
print('number of validation samples: %d' % len(validation_index))
wsi_paths = tumor_wsi_paths + normal_wsi_paths
print(len(wsi_paths))
features_file_train = open(f_train, 'w')
features_file_validation = open(f_validation, 'w')
wr_train = csv.writer(features_file_train, quoting=csv.QUOTE_NONNUMERIC)
wr_validation = csv.writer(features_file_validation, quoting=csv.QUOTE_NONNUMERIC)
wr_train.writerow(utils.heatmap_feature_names)
wr_validation.writerow(utils.heatmap_feature_names)
index = 0
for wsi_path in wsi_paths:
wsi_name = utils.get_filename_from_path(wsi_path)
# print('extracting features for: %s' % wsi_name)
heatmap_prob_path = glob.glob(
os.path.join(utils.HEAT_MAP_DIR, '*%s*%s' % (wsi_name, heatmap_prob_name_postfix_first_model)))
# print(heatmap_prob_path)
image_open = wsi_ops.get_image_open(wsi_path)
heatmap_prob = cv2.imread(heatmap_prob_path[0])
if heatmap_prob_name_postfix_second_model is not None:
heatmap_prob_path_second_model = glob.glob(
os.path.join(utils.HEAT_MAP_DIR, '*%s*%s' % (wsi_name, heatmap_prob_name_postfix_second_model)))
heatmap_prob_second_model = cv2.imread(heatmap_prob_path_second_model[0])
for row in range(heatmap_prob.shape[0]):
for col in range(heatmap_prob.shape[1]):
if heatmap_prob[row, col, 0] >= 0.90 * 255 and heatmap_prob_second_model[row, col, 0] < 0.20 * 255:
heatmap_prob[row, col, :] = heatmap_prob_second_model[row, col, :]
features = extract_features(heatmap_prob, image_open)
if 'umor' in wsi_name:
features += [1]
else:
features += [0]
print(features)
if index in validation_index:
wr_validation.writerow(features)
else:
wr_train.writerow(features)
index += 1
def extract_features_first_heatmap():
# extract_features_train_validation('_prob.png', None, utils.HEATMAP_FEATURE_CSV_TRAIN,
# utils.HEATMAP_FEATURE_CSV_VALIDATION)
# extract_features_train_all('_prob.png', None, utils.HEATMAP_FEATURE_CSV_TRAIN_ALL)
extract_features_test('_prob.png', None, utils.HEATMAP_FEATURE_CSV_TEST)
def extract_features_both_heatmap():
# extract_features_train_validation('_prob.png', '_prob_%s.png' % utils.SECOND_HEATMAP_MODEL,
# utils.HEATMAP_FEATURE_CSV_TRAIN_SECOND_MODEL,
# utils.HEATMAP_FEATURE_CSV_VALIDATION_SECOND_MODEL)
# extract_features_train_all('_prob.png', '_prob_%s.png' % utils.SECOND_HEATMAP_MODEL,
# utils.HEATMAP_FEATURE_CSV_TRAIN_ALL_SECOND_MODEL)
extract_features_test('_prob.png', '_prob_%s.png' % utils.SECOND_HEATMAP_MODEL,
utils.HEATMAP_FEATURE_CSV_TEST_SECOND_MODEL)
if __name__ == '__main__':
wsi_ops = WSIOps()
# extract_features_first_heatmap()
extract_features_both_heatmap()