forked from wikke/Tianchi-Medical-LungTumorDetect
-
Notifications
You must be signed in to change notification settings - Fork 0
/
model_UNet.py
173 lines (130 loc) · 7.3 KB
/
model_UNet.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
from keras.models import Model
from keras.layers import Input, Conv3D, MaxPooling3D, UpSampling3D, concatenate
from keras.callbacks import Callback
from keras.optimizers import Adam
from keras import backend as K
from config import *
from generators import get_seg_batch
from skimage import morphology, measure, segmentation
from visual_utils import plot_slices, plot_comparison
import numpy as np
SMOOTH = 1.0
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_loss(y_true, y_pred):
return 1-dice_coef(y_true, y_pred)
def metrics_true_sum(y_true, y_pred):
return K.sum(y_true)
def metrics_pred_sum(y_true, y_pred):
return K.sum(y_pred)
def metrics_pred_max(y_true, y_pred):
return K.max(y_pred)
def metrics_pred_min(y_true, y_pred):
return K.min(y_pred)
def metrics_pred_mean(y_true, y_pred):
return K.mean(y_pred)
def do_evaluate(model):
print('Model evaluating')
X, y_true = next(get_seg_batch(1, from_train=False, random_choice=True))
y_pred = model.predict(X)
X, y_true, y_pred = X[0,:,:,:,0], y_true[0,:,:,:,0], y_pred[0,:,:,:,0]
intersection = y_true * y_pred
recall = (np.sum(intersection) + SMOOTH) / (np.sum(y_true) + SMOOTH)
precision = (np.sum(intersection) + SMOOTH) / (np.sum(y_pred) + SMOOTH)
print('Average recall {:.4f}, precision {:.4f}'.format(recall, precision))
for threshold in range(0, 10, 2):
threshold = threshold / 10.0
pred_mask = (y_pred > threshold).astype(np.uint8)
intersection = y_true * pred_mask
recall = (np.sum(intersection) + SMOOTH) / (np.sum(y_true) + SMOOTH)
precision = (np.sum(intersection) + SMOOTH) / (np.sum(y_pred) + SMOOTH)
print("Threshold {}: recall {:.4f}, precision {:.4f}".format(threshold, recall, precision))
regions = measure.regionprops(measure.label(y_pred))
print('Num of pred regions {}'.format(len(regions)))
if DEBUG_PLOT_WHEN_EVALUATING_SEG:
plot_comparison(X, y_true, y_pred)
plot_slices(X)
plot_slices(y_true)
plot_slices(y_pred)
class UNetEvaluator(Callback):
def __init__(self):
self.counter = 0
def on_epoch_end(self, epoch, logs=None):
self.counter += 1
if self.counter % TRAIN_SEG_EVALUATE_FREQ == 0:
do_evaluate(self.model)
def get_unet():
return get_simplified_unet() if USE_SIMPLIFIED_UNET else get_full_unet()
def get_simplified_unet():
inputs = Input((INPUT_WIDTH, INPUT_HEIGHT, INPUT_DEPTH, INPUT_CHANNEL))
conv1 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(inputs)
conv1 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(conv1)
pool1 = MaxPooling3D(pool_size=(2, 2, 2))(conv1)
conv2 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(pool1)
conv2 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(conv2)
pool2 = MaxPooling3D(pool_size=(2, 2, 2))(conv2)
conv3 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(pool2)
conv3 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(conv3)
pool3 = MaxPooling3D(pool_size=(2, 2, 2))(conv3)
conv4 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(pool3)
conv4 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(conv4)
# pool4 = MaxPooling3D(pool_size=(2, 2, 2))(conv4)
# conv5 = Conv3D(512, (3, 3, 3), activation='relu', padding='same')(pool4)
# conv5 = Conv3D(512, (3, 3, 3), activation='relu', padding='same')(conv5)
# up6 = concatenate([UpSampling3D(size=(2, 2, 2))(conv5), conv4], axis=-1)
# conv6 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(up6)
# conv6 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(conv6)
# up7 = concatenate([UpSampling3D(size=(2, 2, 2))(conv6), conv3], axis=-1)
up7 = concatenate([UpSampling3D(size=(2, 2, 2))(conv4), conv3], axis=-1)
conv7 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(up7)
conv7 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(conv7)
up8 = concatenate([UpSampling3D(size=(2, 2, 2))(conv7), conv2], axis=-1)
conv8 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(up8)
conv8 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(conv8)
up9 = concatenate([UpSampling3D(size=(2, 2, 2))(conv8), conv1], axis=-1)
conv9 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(up9)
conv9 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(conv9)
conv10 = Conv3D(OUTPUT_CHANNEL, (1, 1, 1), activation='sigmoid')(conv9)
model = Model(inputs=inputs, outputs=conv10)
model.compile(optimizer=Adam(lr=TRAIN_SEG_LEARNING_RATE), loss=dice_coef_loss,
metrics=[dice_coef, metrics_true_sum, metrics_pred_sum, metrics_pred_max, metrics_pred_min, metrics_pred_mean])
return model
def get_full_unet():
inputs = Input((INPUT_WIDTH, INPUT_HEIGHT, INPUT_DEPTH, INPUT_CHANNEL))
conv1 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(inputs)
conv1 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(conv1)
pool1 = MaxPooling3D(pool_size=(2, 2, 2))(conv1)
conv2 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(pool1)
conv2 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(conv2)
pool2 = MaxPooling3D(pool_size=(2, 2, 2))(conv2)
conv3 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(pool2)
conv3 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(conv3)
pool3 = MaxPooling3D(pool_size=(2, 2, 2))(conv3)
conv4 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(pool3)
conv4 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(conv4)
pool4 = MaxPooling3D(pool_size=(2, 2, 2))(conv4)
conv5 = Conv3D(512, (3, 3, 3), activation='relu', padding='same')(pool4)
conv5 = Conv3D(512, (3, 3, 3), activation='relu', padding='same')(conv5)
up6 = concatenate([UpSampling3D(size=(2, 2, 2))(conv5), conv4], axis=-1)
conv6 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(up6)
conv6 = Conv3D(256, (3, 3, 3), activation='relu', padding='same')(conv6)
up7 = concatenate([UpSampling3D(size=(2, 2, 2))(conv6), conv3], axis=-1)
conv7 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(up7)
conv7 = Conv3D(128, (3, 3, 3), activation='relu', padding='same')(conv7)
up8 = concatenate([UpSampling3D(size=(2, 2, 2))(conv7), conv2], axis=-1)
conv8 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(up8)
conv8 = Conv3D(64, (3, 3, 3), activation='relu', padding='same')(conv8)
up9 = concatenate([UpSampling3D(size=(2, 2, 2))(conv8), conv1], axis=-1)
conv9 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(up9)
conv9 = Conv3D(32, (3, 3, 3), activation='relu', padding='same')(conv9)
conv10 = Conv3D(OUTPUT_CHANNEL, (1, 1, 1), activation='sigmoid')(conv9)
model = Model(inputs=inputs, outputs=conv10)
model.compile(optimizer=Adam(lr=TRAIN_SEG_LEARNING_RATE), loss=dice_coef_loss,
metrics=[dice_coef, metrics_true_sum, metrics_pred_sum, metrics_pred_max, metrics_pred_min, metrics_pred_mean])
return model
if __name__ == '__main__':
model = get_unet()
model.summary()