forked from jocicmarko/ultrasound-nerve-segmentation
/
train_shiftbn_sgd.py
272 lines (220 loc) · 10.1 KB
/
train_shiftbn_sgd.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
from __future__ import print_function
'''
seed = 1024,10 epoch
seed = 1024+1,40 epoch
datagen = ImageDataGenerator(
featurewise_center=False, # set input mean to 0 over the dataset
samplewise_center=False, # set each sample mean to 0
featurewise_std_normalization=False, # divide inputs by std of the dataset
samplewise_std_normalization=False, # divide each input by its std
zca_whitening=False, # apply ZCA whitening
rotation_range=45, # randomly rotate images in the range (degrees, 0 to 180)
width_shift_range=0.0, # randomly shift images horizontally (fraction of total width)
height_shift_range=0.0, # randomly shift images vertically (fraction of total height)
horizontal_flip=True, # randomly flip images
vertical_flip=False)
'''
import cv2
import numpy as np
from keras.models import Model
from keras.layers import Input, merge, Convolution2D, MaxPooling2D, UpSampling2D
from keras.layers.normalization import BatchNormalization
from keras.layers.advanced_activations import PReLU,LeakyReLU,ELU,SReLU
from keras.optimizers import Adam,SGD#,Nadam
from keras.preprocessing.image import ImageDataGenerator
from keras.callbacks import ModelCheckpoint, LearningRateScheduler
from keras import backend as K
from keras.utils.visualize_util import plot
from sklearn.cross_validation import train_test_split
from sklearn.cross_validation import KFold,StratifiedKFold
from data import load_train_data, load_test_data
seed = 1024
np.random.seed(seed)
img_rows = 64#*2
img_cols = 80#*2
smooth = 1.
def dice_coef(y_true, y_pred):
y_true_f = K.flatten(y_true)
y_pred_f = K.flatten(y_pred)
intersection = K.sum(y_true_f * y_pred_f)
return (2. * intersection + smooth) / (K.sum(y_true_f) + K.sum(y_pred_f) + smooth)
def dice_coef(y_true, y_pred):
y_true_f = K.flatten(y_true)
y_pred_f = K.flatten(y_pred)
intersection = K.sum(y_true_f * y_pred_f)
return (2. * intersection) / (K.sum(y_true_f) + K.sum(y_pred_f))
def dice_coef_loss(y_true, y_pred):
return -dice_coef(y_true, y_pred)
def mask_not_blank(mask):
return sum(mask.flatten()) > 0
def get_unet():
inputs = Input((1, img_rows, img_cols))
conv1 = Convolution2D(32, 3, 3, border_mode='same')(inputs)
conv1 = SReLU()(conv1)
conv1 = Convolution2D(32, 3, 3, border_mode='same')(conv1)
conv1 = SReLU()(conv1)
conv1 = BatchNormalization(axis=1)(conv1)
pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)
conv2 = Convolution2D(64, 3, 3, border_mode='same')(pool1)
conv2 = SReLU()(conv2)
conv2= Convolution2D(64, 3, 3, border_mode='same')(conv2)
conv2 = SReLU()(conv2)
conv2 = BatchNormalization(axis=1)(conv2)
pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
conv3 = Convolution2D(128, 3, 3, border_mode='same')(pool2)
conv3 = SReLU()(conv3)
conv3 = Convolution2D(128, 3, 3, border_mode='same')(conv3)
conv3 = SReLU()(conv3)
conv3 = BatchNormalization(axis=1)(conv3)
pool3 = MaxPooling2D(pool_size=(2, 2))(conv3)
conv4 = Convolution2D(256, 3, 3, border_mode='same')(pool3)
conv4 = SReLU()(conv4)
conv4 = Convolution2D(256, 3, 3, border_mode='same')(conv4)
conv4 = SReLU()(conv4)
conv4 = BatchNormalization(axis=1)(conv4)
pool4 = MaxPooling2D(pool_size=(2, 2))(conv4)
conv5 = Convolution2D(512, 3, 3, border_mode='same')(pool4)
conv5 = SReLU()(conv5)
conv5 = Convolution2D(512, 3, 3, border_mode='same')(conv5)
conv5 = SReLU()(conv5)
conv5 = BatchNormalization(axis=1)(conv5)
up6 = merge([UpSampling2D(size=(2, 2))(conv5), conv4], mode='concat', concat_axis=1)
conv6 = Convolution2D(256, 3, 3, border_mode='same')(up6)
conv6 = SReLU()(conv6)
conv6 = Convolution2D(256, 3, 3, border_mode='same')(conv6)
conv6 = SReLU()(conv6)
conv6 = BatchNormalization(axis=1)(conv6)
up7 = merge([UpSampling2D(size=(2, 2))(conv6), conv3], mode='concat', concat_axis=1)
conv7 = Convolution2D(128, 3, 3, border_mode='same')(up7)
conv7 = SReLU()(conv7)
conv7 = Convolution2D(128, 3, 3, border_mode='same')(conv7)
conv7 = SReLU()(conv7)
conv7 = BatchNormalization(axis=1)(conv7)
up8 = merge([UpSampling2D(size=(2, 2))(conv7), conv2], mode='concat', concat_axis=1)
conv8 = Convolution2D(64, 3, 3, border_mode='same')(up8)
conv8 = SReLU()(conv8)
conv8 = Convolution2D(64, 3, 3, border_mode='same')(conv8)
conv8 = SReLU()(conv8)
conv8 = BatchNormalization(axis=1)(conv8)
up9 = merge([UpSampling2D(size=(2, 2))(conv8), conv1], mode='concat', concat_axis=1)
conv9 = Convolution2D(32, 3, 3, border_mode='same')(up9)
conv9 = SReLU()(conv9)
conv9 = Convolution2D(32, 3, 3, border_mode='same')(conv9)
conv9 = SReLU()(conv9)
conv9 = BatchNormalization(axis=1)(conv9)
'''
output
'''
conv10 = Convolution2D(1, 1, 1, activation='sigmoid')(conv9)
model = Model(input=inputs, output=conv10)
# sgd = sgd()
sgd = SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(optimizer=sgd, loss=dice_coef_loss, metrics=[dice_coef])
return model
def preprocess(imgs):
imgs_p = np.ndarray((imgs.shape[0], imgs.shape[1], img_rows, img_cols), dtype=np.uint8)
for i in range(imgs.shape[0]):
imgs_p[i, 0] = cv2.resize(imgs[i, 0], (img_cols, img_rows), interpolation=cv2.INTER_CUBIC)
return imgs_p
def train_and_predict():
print('-'*30)
print('Loading and preprocessing train data...')
print('-'*30)
imgs_train, imgs_mask_train = load_train_data()
imgs_train = preprocess(imgs_train)
imgs_mask_train = preprocess(imgs_mask_train)
imgs_train = imgs_train.astype('float32')
mean = np.mean(imgs_train) # mean for data centering
std = np.std(imgs_train) # std for data normalization
imgs_train -= mean
imgs_train /= std
imgs_mask_train = imgs_mask_train.astype('float32')
imgs_mask_train /= 255. # scale masks to [0, 1]
y_bin = np.array([mask_not_blank(mask) for mask in imgs_mask_train ])
# X_train,X_test,y_train,y_test = train_test_split(imgs_train,imgs_mask_train,test_size=0.2,random_state=seed)
skf = StratifiedKFold(y_bin, n_folds=10, shuffle=True, random_state=seed)
for ind_tr, ind_te in skf:
X_train = imgs_train[ind_tr]
X_test = imgs_train[ind_te]
y_train = imgs_mask_train[ind_tr]
y_test = imgs_mask_train[ind_te]
break
X_train_flip = X_train[:,:,:,::-1]
y_train_flip = y_train[:,:,:,::-1]
X_train = np.concatenate((X_train,X_train_flip),axis=0)
y_train = np.concatenate((y_train,y_train_flip),axis=0)
X_train_flip = X_train[:,:,::-1,:]
y_train_flip = y_train[:,:,::-1,:]
X_train = np.concatenate((X_train,X_train_flip),axis=0)
y_train = np.concatenate((y_train,y_train_flip),axis=0)
imgs_train = X_train
imgs_valid = X_test
imgs_mask_train = y_train
imgs_mask_valid = y_test
print('-'*30)
print('Creating and compiling model...')
print('-'*30)
model = get_unet()
model_name = 'unet_seed_1024_epoch_20_aug_64_80_shiftbn_sgd_srelu_plus10.hdf5'
model_checkpoint = ModelCheckpoint('E:\\UltrasoundNerve\\'+model_name, monitor='loss', save_best_only=True)
plot(model, to_file='E:\\UltrasoundNerve\\%s.png'%model_name.replace('.hdf5',''),show_shapes=True)
print('-'*30)
print('Fitting model...')
print('-'*30)
augmentation=False
batch_size=128
nb_epoch=10
load_model=True
use_all_data = False
if use_all_data:
imgs_train = np.concatenate((imgs_train,imgs_valid),axis=0)
imgs_mask_train = np.concatenate((imgs_mask_train,imgs_mask_valid),axis=0)
if load_model:
model.load_weights('E:\\UltrasoundNerve\\'+model_name)
if not augmentation:
# model.fit(imgs_train, imgs_mask_train, batch_size=batch_size, nb_epoch=nb_epoch, verbose=1, shuffle=True,
# callbacks=[model_checkpoint],
# validation_data=[imgs_valid,imgs_mask_valid]
# )
pass
else:
datagen = ImageDataGenerator(
featurewise_center=False, # set input mean to 0 over the dataset
samplewise_center=False, # set each sample mean to 0
featurewise_std_normalization=False, # divide inputs by std of the dataset
samplewise_std_normalization=False, # divide each input by its std
zca_whitening=False, # apply ZCA whitening
rotation_range=45, # randomly rotate images in the range (degrees, 0 to 180)
width_shift_range=0.0, # randomly shift images horizontally (fraction of total width)
height_shift_range=0.0, # randomly shift images vertically (fraction of total height)
horizontal_flip=True, # randomly flip images
vertical_flip=False) # randomly flip images
# compute quantities required for featurewise normalization
# (std, mean, and principal components if ZCA whitening is applied)
datagen.fit(imgs_train)
# fit the model on the batches generated by datagen.flow()
model.fit_generator(datagen.flow(imgs_train, imgs_mask_train,
batch_size=batch_size),
samples_per_epoch=imgs_train.shape[0],
nb_epoch=nb_epoch,
callbacks=[model_checkpoint],
validation_data=(imgs_valid,imgs_mask_valid))
print('-'*30)
print('Loading and preprocessing test data...')
print('-'*30)
imgs_test, imgs_id_test = load_test_data()
imgs_test = preprocess(imgs_test)
imgs_test = imgs_test.astype('float32')
imgs_test -= mean
imgs_test /= std
print('-'*30)
print('Loading saved weights...')
print('-'*30)
model.load_weights('E:\\UltrasoundNerve\\'+model_name)
print('-'*30)
print('Predicting masks on test data...')
print('-'*30)
imgs_mask_test = model.predict(imgs_test, verbose=1)
np.save('imgs_mask_test.npy', imgs_mask_test)
if __name__ == '__main__':
train_and_predict()