-
Notifications
You must be signed in to change notification settings - Fork 0
/
create_data.py
421 lines (354 loc) · 14.1 KB
/
create_data.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
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
# coding = utf-8
# -*- coding:utf-8 -*-
import os
import re
import shutil
from collections import defaultdict
from math import *
import random
import fire
import numpy as np
import pathlib
import scipy.spatial.transform as f
import torch.nn as nn
import torch.utils.data
from torch.utils.data import Dataset
from pcdet.utils import common_utils
from pcdet.datasets.augmentor.data_augmentor import DataAugmentor
from pcdet.datasets.processor.data_processor import DataProcessor
from pcdet.datasets.processor.point_feature_encoder import PointFeatureEncoder
from sklearn.utils import class_weight
def read_one_xyz(filename):
xyz = []
with open(filename, 'r') as f:
content = f.read()
contact = content.split('\n')
for line in contact:
if line == '' or line.isdigit():
continue
else:
atom = line.split()
xyzitem = [atom[0], atom[1], atom[2]]
xyz.append(xyzitem)
return np.array(xyz).astype(float)
def getfiles(dirPath, fileType):
fileList = []
files = os.listdir(dirPath)
files.sort()
pattern = re.compile('.*/' + fileType)
for f in files:
if os.path.isdir(dirPath + '/' + f):
getfiles(dirPath + '/' + f, fileType)
elif os.path.isfile(dirPath + '/' + f):
matches = pattern.match(f)
if matches is not None:
fileList.append(dirPath + '/' + matches.group())
# else:
# fileList.append(dirPath + '/invalid')
return fileList
def points_cloud_processing(translation=True):
path = '/home/s2020153/cardiac/DATASET/ACDC_ALL_FRAMES_SEG_FINAL'
input_files = os.listdir(path)
fType = '.xyz'
lv = getfiles(path + os.sep + input_files[0], fType)
myo = getfiles(path + os.sep + input_files[1], fType)
rv = getfiles(path + os.sep + input_files[2], fType)
for i in range(len(lv)):
lv_in = read_one_xyz(lv[i])
myo_in = read_one_xyz(myo[i])
rv_in = read_one_xyz(rv[i])
if translation:
lv_centre = np.average(lv_in, axis=0)
myo_centre = np.average(myo_in, axis=0)
rv_centre = np.average(rv_in, axis=0)
myo_offset = lv_centre - myo_centre
myo_in = myo_in + myo_offset
myo_centre = np.average(myo_in, axis=0)
mid_point = (rv_centre + lv_centre) / 2
translation_matrix = -mid_point
# print(translation_matrix)
lv_in = lv_in + translation_matrix
rv_in = rv_in + translation_matrix
myo_in = myo_in + translation_matrix
lv_centre2 = np.average(lv_in, axis=0)
myo_centre2 = np.average(myo_in, axis=0)
rv_centre2 = np.average(rv_in, axis=0)
lv_rv_vec = lv_centre2 - rv_centre2
unit_lv_rv_vec = lv_rv_vec / np.linalg.norm(lv_rv_vec)
x_vect = np.array([1, 0, 0])
z_vect = np.array([0, 0, 1])
dot_product = np.dot(unit_lv_rv_vec, x_vect)
angle = np.arccos(dot_product)
# angle = np.rad2deg(angle)
# print(angle)
rot_angle = -(np.pi / 4 - angle)
rot_vect = rot_angle * z_vect
rotation = f.Rotation.from_rotvec(rot_vect)
lv_in = rotation.apply(lv_in)
myo_in = rotation.apply(myo_in)
rv_in = rotation.apply(rv_in)
lv_out = np.array([np.append(ls, [int(lv[i][-7:-4]), 1]) for ls in lv_in])
myo_out = np.array([np.append(ls, [int(myo[i][-7:-4]), 2]) for ls in myo_in])
rv_out = np.array([np.append(ls, [int(rv[i][-7:-4]), 3]) for ls in rv_in])
lv_out = random.sample(list(lv_out), 5000)
myo_out = random.sample(list(myo_out), 10000)
rv_out = random.sample(list(rv_out), 4000)
output = np.concatenate((lv_out, myo_out, rv_out))
out_save_file_tr = '/home/s2020153/cardiac/DATASET/training' + os.sep + lv[i][-18:-4] + '.bin'
output.tofile(out_save_file_tr)
def points_generator(name):
root_path = r'/home/s2020153/cardiac/DATASET/' + name
ls = os.listdir(root_path)
ls.sort()
length = len(ls)
count = 1
k = 1
pc = np.fromfile(os.path.join(root_path, ls[0]))
print(ls[0])
set_path = pathlib.Path(r'/home/s2020153/cardiac/DATASET/{}_set'.format(name))
set_path.mkdir(parents=True, exist_ok=True)
while count < length and k < 10:
pc2 = np.fromfile(os.path.join(root_path, ls[count]))
pc = np.concatenate((pc, pc2))
print(ls[count])
k += 1
count += 1
if k == 10:
out_save_path = r'/home/s2020153/cardiac/DATASET/{}_set'.format(name) + os.sep + ls[count - 1][:-8] + '.bin'
pc.tofile(out_save_path)
if count < length:
pc = np.fromfile(os.path.join(root_path, ls[count]))
print('------------------------------------------------------------------')
print(ls[count])
count += 1
k = 1
def points_reader(name, path):
root_path = r'/home/s2020153/cardiac/DATASET/' + name
pc = np.fromfile(os.path.join(root_path, path))
pc = pc.reshape(-1, 5)
return pc
def get_file_names(name):
root_path = '/home/s2020153/cardiac/DATASET/' + name
names = os.listdir(root_path)
names.sort()
return names
def get_labels():
root_path = r'/home/s2020153/cardiac/cardiac_pillars/labels'
files = os.listdir(root_path)
files.sort()
labels = []
for fi in files:
contents = []
with open(root_path + os.sep + fi) as txt:
for line in txt:
line = line.strip('\n')
line = line.rstrip()
contents.append(line)
labels.append(contents[2][7:])
return labels
def get_ED_ES_from_file(name, training=True):
contents = []
if training:
root_path = r'/home/s2020153/cardiac/cardiac_pillars/labels'
else:
root_path = r'/home/s2020153/cardiac/cardiac_pillars/test_info'
with open(root_path + os.sep + name + '.txt') as txt:
for line in txt:
line = line.strip('\n')
line = line.rstrip()
contents.append(line)
try:
return int(contents[0][-1]), int(contents[1][-2:])
except ValueError:
return int(contents[0][-1]), int(contents[1][-1])
def set_picked_frames(d, s):
ls = [d - 1, s - 1]
if s < 10:
for i in range(d, s - 1):
ls.append(i)
while len(ls) != 10:
ls.sort()
ls.append(ls[-1] + 1)
return [i + 1 for i in ls]
elif d > 1:
for i in range(d, s - 1):
ls.append(i)
while len(ls) != 10:
ls.append(ls[0] - 1)
ls.sort()
return [i + 1 for i in ls]
else:
for i in range(d, s - 1, 2):
ls.append(i)
if s % 2 == 0:
for j in range(s - 2, d, -2):
ls.append(j)
if len(ls) == 10:
break
if s % 2 != 0:
for j in range(s - 3, d, -2):
ls.append(j)
if len(ls) == 10:
break
return [i + 1 for i in ls]
def get_label_files_names(training):
if training:
root_path = r'/home/s2020153/cardiac/cardiac_pillars/labels'
else:
root_path = r'/home/s2020153/cardiac/cardiac_pillars/test_info'
files = os.listdir(root_path)
files.sort()
return [fi[:-4] for fi in files]
def train_val_split(file_name, training=True):
if training:
path = r'/home/s2020153/cardiac/DATASET/training'
else:
path = r'/home/s2020153/cardiac/DATASET/testing'
files = get_file_names(file_name)
names = get_label_files_names(training)
for i in names:
ed, es = get_ED_ES_from_file(i, training)
frames = set_picked_frames(ed, es)
for j in files:
if (i in j) and (int(j[-7:-4]) not in frames):
os.remove(path + os.sep + j)
def get_acdc_features(name):
path = r'/home/s2020153/cardiac/cardiac_pillars/manual_features/acdc_{}.txt'.format(name)
contents = []
with open(path) as file:
for line in file:
line = line.strip('\n')
line = line.rstrip()
contents.append(line)
contents = [i.split(' ') for i in contents]
thick_dict = {}
for ls in contents:
thick_dict.update({ls[0]: np.array(ls[1:], dtype=float)})
return thick_dict
class CardiacDataset(Dataset):
def __init__(self, train_or_val, transform=None, target_transform=None, loader=None, root_path=None, config=None,
training=True):
super(CardiacDataset, self).__init__()
if config is None:
return
# acdc_thickness = get_acdc_features('thickness')
# acdc_volume = get_acdc_features('volume')
class_names = config.CLASS_NAMES
origin_labels = get_labels()
points = get_file_names(train_or_val)
if training:
labels = [class_names.index(origin_labels[int(i[-7:-4]) - 1]) for i in points
if int(i[-7:-4]) <= 100]
labels.extend([class_names.index(origin_labels[int(j[-7:-4]) - 51]) for j in points
if int(j[-7:-4]) >= 151])
labels = np.array(labels)
else:
labels = []
point_cloud_range = np.array([0, -39.68, -3, 69.12, 39.68, 1])
voxel_size = np.array([0.16, 0.16, 4])
grid_size = (point_cloud_range[3:6] - point_cloud_range[0:3]) / voxel_size
# self.acdc_thickness = acdc_thickness
# self.acdc_volume = acdc_volume
self.train_or_val = train_or_val
self.config = config
self.root_path = root_path
self.class_names = class_names
self.training = training
self.voxel_size = voxel_size
self.point_cloud_range = point_cloud_range
self.grid_size = np.round(grid_size).astype(np.int64)
self.points = points
self.labels = labels
self.loader = loader
self.transform = transform
self.target_transform = target_transform
self.point_feature_encoder = PointFeatureEncoder(self.config.DATA_CONFIG.POINT_FEATURE_ENCODING,
point_cloud_range=self.point_cloud_range)
self.data_augmentor = DataAugmentor(self.root_path, self.config.DATA_CONFIG.DATA_AUGMENTOR,
self.class_names, logger=None) if self.training else None
self.data_processor = DataProcessor(
self.config.DATA_CONFIG.DATA_PROCESSOR, point_cloud_range=self.point_cloud_range, training=self.training
)
def get_labels(self):
return self.labels
def prepare_data(self, data_dict):
"""
Args:
data_dict:
points: optional, (N, 3 + C_in)
label: optional, (1), string
...
Returns:
data_dict:
points: (N, 3 + C_in)
gt_names: optional, (1), string
use_lead_xyz: bool
voxels: optional (num_voxels, max_points_per_voxel, 3 + C)
voxel_coords: optional (num_voxels, 3)
voxel_num_points: optional (num_voxels)
...
"""
if self.training:
data_dict = self.data_augmentor.forward(
data_dict={
**data_dict,
}
)
if data_dict.get('points', None) is not None:
data_dict = self.point_feature_encoder.forward(data_dict)
data_dict = self.data_processor.forward(
data_dict=data_dict
)
return data_dict
@staticmethod
def collate_batch(batch_list, _unused=False):
data_dict = defaultdict(list)
for cur_sample in batch_list:
for key, val in cur_sample.items():
data_dict[key].append(val)
batch_size = len(batch_list)
ret = {}
for key, val in data_dict.items():
try:
# voxels: optional (num_voxels, max_points_per_voxel, 3 + C)
# voxel_coords: optional (num_voxels, 3)
# voxel_num_points: optional (num_voxels)
if key in ['voxels', 'voxel_num_points']:
ret[key] = np.concatenate(val, axis=0)
elif key in ['points', 'voxel_coords']:
coors = []
for i, coor in enumerate(val):
coor_pad = np.pad(coor, ((0, 0), (1, 0)), mode='constant', constant_values=i)
coors.append(coor_pad)
ret[key] = np.concatenate(coors, axis=0) # (B, N, 4)
else:
ret[key] = np.stack(val, axis=0)
except:
print('Error in collate_batch: key=%s' % key)
raise TypeError
ret['batch_size'] = batch_size
return ret
def __getitem__(self, item):
points_name = self.points[item]
if self.training:
label = np.array(self.labels[item])
else:
label = []
# thick = self.acdc_thickness[points_name[:-4]]
# volume = self.acdc_volume[points_name[:-4]]
pc = self.loader(self.train_or_val, points_name)
if self.transform is not None:
pc = self.transform(pc)
input_dict = {
'name': points_name[:-4],
'points': pc,
'label': label,
# 'thickness': thick,
# 'volume': volume
}
data_dict = self.prepare_data(input_dict)
return data_dict
def __len__(self):
return len(self.points)
if __name__ == '__main__':
fire.Fire()