-
Notifications
You must be signed in to change notification settings - Fork 186
/
models.py
318 lines (265 loc) · 15.2 KB
/
models.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
import numpy as np
import os
import lpips.lpips_tf as lpips_tf
import tensorflow as tf
import utils
from tensorflow import keras
from tensorflow.python.keras.models import *
from tensorflow.python.keras.layers import *
from stn import spatial_transformer_network as stn_transformer
class StegaStampEncoder(Layer):
def __init__(self, height, width):
super(StegaStampEncoder, self).__init__()
self.secret_dense = Dense(7500, activation='relu', kernel_initializer='he_normal')
self.conv1 = Conv2D(32, 3, activation='relu', padding='same', kernel_initializer='he_normal')
self.conv2 = Conv2D(32, 3, activation='relu', strides=2, padding='same', kernel_initializer='he_normal')
self.conv3 = Conv2D(64, 3, activation='relu', strides=2, padding='same', kernel_initializer='he_normal')
self.conv4 = Conv2D(128, 3, activation='relu', strides=2, padding='same', kernel_initializer='he_normal')
self.conv5 = Conv2D(256, 3, activation='relu', strides=2, padding='same', kernel_initializer='he_normal')
self.up6 = Conv2D(128, 2, activation='relu', padding='same', kernel_initializer='he_normal')
self.conv6 = Conv2D(128, 3, activation='relu', padding='same', kernel_initializer='he_normal')
self.up7 = Conv2D(64, 2, activation='relu', padding='same', kernel_initializer='he_normal')
self.conv7 = Conv2D(64, 3, activation='relu', padding='same', kernel_initializer='he_normal')
self.up8 = Conv2D(32, 2, activation='relu', padding='same', kernel_initializer='he_normal')
self.conv8 = Conv2D(32, 3, activation='relu', padding='same', kernel_initializer='he_normal')
self.up9 = Conv2D(32, 2, activation='relu', padding='same', kernel_initializer='he_normal')
self.conv9 = Conv2D(32, 3, activation='relu', padding='same', kernel_initializer='he_normal')
self.conv10 = Conv2D(32, 3, activation='relu', padding='same', kernel_initializer='he_normal')
self.residual = Conv2D(3, 1, activation=None, padding='same', kernel_initializer='he_normal')
def call(self, inputs):
secret, image = inputs
secret = secret - .5
image = image - .5
secret = self.secret_dense(secret)
secret = Reshape((50, 50, 3))(secret)
secret_enlarged = UpSampling2D(size=(8,8))(secret)
inputs = concatenate([secret_enlarged, image], axis=-1)
conv1 = self.conv1(inputs)
conv2 = self.conv2(conv1)
conv3 = self.conv3(conv2)
conv4 = self.conv4(conv3)
conv5 = self.conv5(conv4)
up6 = self.up6(UpSampling2D(size=(2,2))(conv5))
merge6 = concatenate([conv4,up6], axis=3)
conv6 = self.conv6(merge6)
up7 = self.up7(UpSampling2D(size=(2,2))(conv6))
merge7 = concatenate([conv3,up7], axis=3)
conv7 = self.conv7(merge7)
up8 = self.up8(UpSampling2D(size=(2,2))(conv7))
merge8 = concatenate([conv2,up8], axis=3)
conv8 = self.conv8(merge8)
up9 = self.up9(UpSampling2D(size=(2,2))(conv8))
merge9 = concatenate([conv1,up9,inputs], axis=3)
conv9 = self.conv9(merge9)
conva = self.conv9(merge9)
conv10 = self.conv10(conv9)
residual = self.residual(conv9)
return residual
class StegaStampDecoder(Layer):
def __init__(self, secret_size, height, width):
super(StegaStampDecoder, self).__init__()
self.height = height
self.width = width
self.stn_params = Sequential([
Conv2D(32, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(64, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(128, (3, 3), strides=2, activation='relu', padding='same'),
Flatten(),
Dense(128, activation='relu')
])
initial = np.array([[1., 0, 0], [0, 1., 0]])
initial = initial.astype('float32').flatten()
self.W_fc1 = tf.Variable(tf.zeros([128, 6]), name='W_fc1')
self.b_fc1 = tf.Variable(initial_value=initial, name='b_fc1')
self.decoder = Sequential([
Conv2D(32, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(32, (3, 3), activation='relu', padding='same'),
Conv2D(64, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(64, (3, 3), activation='relu', padding='same'),
Conv2D(64, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(128, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(128, (3, 3), strides=2, activation='relu', padding='same'),
Flatten(),
Dense(512, activation='relu'),
Dense(secret_size)
])
def call(self, image):
image = image - .5
stn_params = self.stn_params(image)
x = tf.matmul(stn_params, self.W_fc1) + self.b_fc1
transformed_image = stn_transformer(image, x, [self.height, self.width, 3])
return self.decoder(transformed_image)
class Discriminator(Layer):
def __init__(self):
super(Discriminator, self).__init__()
self.model = Sequential([
Conv2D(8, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(16, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(32, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(64, (3, 3), strides=2, activation='relu', padding='same'),
Conv2D(1, (3, 3), activation=None, padding='same')
])
def call(self, image):
x = image - .5
x = self.model(x)
output = tf.reduce_mean(x)
return output, x
def transform_net(encoded_image, args, global_step):
sh = tf.shape(encoded_image)
ramp_fn = lambda ramp : tf.minimum(tf.to_float(global_step) / ramp, 1.)
rnd_bri = ramp_fn(args.rnd_bri_ramp) * args.rnd_bri
rnd_hue = ramp_fn(args.rnd_hue_ramp) * args.rnd_hue
rnd_brightness = utils.get_rnd_brightness_tf(rnd_bri, rnd_hue, args.batch_size)
jpeg_quality = 100. - tf.random.uniform([]) * ramp_fn(args.jpeg_quality_ramp) * (100.-args.jpeg_quality)
jpeg_factor = tf.cond(tf.less(jpeg_quality, 50), lambda: 5000. / jpeg_quality, lambda: 200. - jpeg_quality * 2) / 100. + .0001
rnd_noise = tf.random.uniform([]) * ramp_fn(args.rnd_noise_ramp) * args.rnd_noise
contrast_low = 1. - (1. - args.contrast_low) * ramp_fn(args.contrast_ramp)
contrast_high = 1. + (args.contrast_high - 1.) * ramp_fn(args.contrast_ramp)
contrast_params = [contrast_low, contrast_high]
rnd_sat = tf.random.uniform([]) * ramp_fn(args.rnd_sat_ramp) * args.rnd_sat
# blur
f = utils.random_blur_kernel(probs=[.25,.25], N_blur=7,
sigrange_gauss=[1.,3.], sigrange_line=[.25,1.], wmin_line=3)
encoded_image = tf.nn.conv2d(encoded_image, f, [1,1,1,1], padding='SAME')
noise = tf.random_normal(shape=tf.shape(encoded_image), mean=0.0, stddev=rnd_noise, dtype=tf.float32)
encoded_image = encoded_image + noise
encoded_image = tf.clip_by_value(encoded_image, 0, 1)
contrast_scale = tf.random_uniform(shape=[tf.shape(encoded_image)[0]], minval=contrast_params[0], maxval=contrast_params[1])
contrast_scale = tf.reshape(contrast_scale, shape=[tf.shape(encoded_image)[0],1,1,1])
encoded_image = encoded_image * contrast_scale
encoded_image = encoded_image + rnd_brightness
encoded_image = tf.clip_by_value(encoded_image, 0, 1)
encoded_image_lum = tf.expand_dims(tf.reduce_sum(encoded_image * tf.constant([.3,.6,.1]), axis=3), 3)
encoded_image = (1 - rnd_sat) * encoded_image + rnd_sat * encoded_image_lum
encoded_image = tf.reshape(encoded_image, [-1,400,400,3])
if not args.no_jpeg:
encoded_image = utils.jpeg_compress_decompress(encoded_image, rounding=utils.round_only_at_0, factor=jpeg_factor, downsample_c=True)
summaries = [tf.summary.scalar('transformer/rnd_bri', rnd_bri),
tf.summary.scalar('transformer/rnd_sat', rnd_sat),
tf.summary.scalar('transformer/rnd_hue', rnd_hue),
tf.summary.scalar('transformer/rnd_noise', rnd_noise),
tf.summary.scalar('transformer/contrast_low', contrast_low),
tf.summary.scalar('transformer/contrast_high', contrast_high),
tf.summary.scalar('transformer/jpeg_quality', jpeg_quality)]
return encoded_image, summaries
def get_secret_acc(secret_true,secret_pred):
with tf.variable_scope("acc"):
secret_pred = tf.round(tf.sigmoid(secret_pred))
correct_pred = tf.to_int64(tf.shape(secret_pred)[1]) - tf.count_nonzero(secret_pred - secret_true, axis=1)
str_acc = 1.0 - tf.count_nonzero(correct_pred - tf.to_int64(tf.shape(secret_pred)[1])) / tf.size(correct_pred, out_type=tf.int64)
bit_acc = tf.reduce_sum(correct_pred) / tf.size(secret_pred, out_type=tf.int64)
return bit_acc, str_acc
def build_model(encoder,
decoder,
discriminator,
secret_input,
image_input,
l2_edge_gain,
borders,
secret_size,
M,
loss_scales,
yuv_scales,
args,
global_step):
input_warped = tf.contrib.image.transform(image_input, M[:,1,:], interpolation='BILINEAR')
mask_warped = tf.contrib.image.transform(tf.ones_like(input_warped), M[:,1,:], interpolation='BILINEAR')
input_warped += (1-mask_warped) * image_input
residual_warped = encoder((secret_input, input_warped))
encoded_warped = residual_warped + input_warped
residual = tf.contrib.image.transform(residual_warped, M[:,0,:], interpolation='BILINEAR')
if borders == 'no_edge':
encoded_image = image_input + residual
elif borders == 'black':
encoded_image = residual_warped + input_warped
encoded_image = tf.contrib.image.transform(encoded_image, M[:,0,:], interpolation='BILINEAR')
input_unwarped = tf.contrib.image.transform(input_warped, M[:,0,:], interpolation='BILINEAR')
elif borders.startswith('random'):
mask = tf.contrib.image.transform(tf.ones_like(residual), M[:,0,:], interpolation='BILINEAR')
encoded_image = residual_warped + input_warped
encoded_image = tf.contrib.image.transform(encoded_image, M[:,0,:], interpolation='BILINEAR')
input_unwarped = tf.contrib.image.transform(input_warped, M[:,0,:], interpolation='BILINEAR')
ch = 3 if borders.endswith('rgb') else 1
encoded_image += (1-mask) * tf.ones_like(residual) * tf.random.uniform([ch])
elif borders == 'white':
mask = tf.contrib.image.transform(tf.ones_like(residual), M[:,0,:], interpolation='BILINEAR')
encoded_image = residual_warped + input_warped
encoded_image = tf.contrib.image.transform(encoded_image, M[:,0,:], interpolation='BILINEAR')
input_unwarped = tf.contrib.image.transform(input_warped, M[:,0,:], interpolation='BILINEAR')
encoded_image += (1-mask) * tf.ones_like(residual)
elif borders == 'image':
mask = tf.contrib.image.transform(tf.ones_like(residual), M[:,0,:], interpolation='BILINEAR')
encoded_image = residual_warped + input_warped
encoded_image = tf.contrib.image.transform(encoded_image, M[:,0,:], interpolation='BILINEAR')
encoded_image += (1-mask) * tf.manip.roll(image_input, shift=1, axis=0)
if borders == 'no_edge':
D_output_real, _ = discriminator(image_input)
D_output_fake, D_heatmap = discriminator(encoded_image)
else:
D_output_real, _ = discriminator(input_warped)
D_output_fake, D_heatmap = discriminator(encoded_warped)
transformed_image, transform_summaries = transform_net(encoded_image, args, global_step)
decoded_secret = decoder(transformed_image)
bit_acc, str_acc = get_secret_acc(secret_input, decoded_secret)
lpips_loss_op = tf.reduce_mean(lpips_tf.lpips(image_input, encoded_image))
secret_loss_op = tf.losses.sigmoid_cross_entropy(secret_input, decoded_secret)
size = (int(image_input.shape[1]),int(image_input.shape[2]))
gain = 10
falloff_speed = 4 # Cos dropoff that reaches 0 at distance 1/x into image
falloff_im = np.ones(size)
for i in range(int(falloff_im.shape[0]/falloff_speed)):
falloff_im[-i,:] *= (np.cos(4*np.pi*i/size[0]+np.pi)+1)/2
falloff_im[i,:] *= (np.cos(4*np.pi*i/size[0]+np.pi)+1)/2
for j in range(int(falloff_im.shape[1]/falloff_speed)):
falloff_im[:,-j] *= (np.cos(4*np.pi*j/size[0]+np.pi)+1)/2
falloff_im[:,j] *= (np.cos(4*np.pi*j/size[0]+np.pi)+1)/2
falloff_im = 1-falloff_im
falloff_im = tf.convert_to_tensor(falloff_im, dtype=tf.float32)
falloff_im *= l2_edge_gain
encoded_image_yuv = tf.image.rgb_to_yuv(encoded_image)
image_input_yuv = tf.image.rgb_to_yuv(image_input)
im_diff = encoded_image_yuv-image_input_yuv
im_diff += im_diff * tf.expand_dims(falloff_im, axis=[-1])
yuv_loss_op = tf.reduce_mean(tf.square(im_diff), axis=[0,1,2])
image_loss_op = tf.tensordot(yuv_loss_op, yuv_scales, axes=1)
D_loss = D_output_real - D_output_fake
G_loss = D_output_fake
loss_op = loss_scales[0]*image_loss_op + loss_scales[1]*lpips_loss_op + loss_scales[2]*secret_loss_op
if not args.no_gan:
loss_op += loss_scales[3]*G_loss
summary_op = tf.summary.merge([
tf.summary.scalar('bit_acc', bit_acc, family='train'),
tf.summary.scalar('str_acc', str_acc, family='train'),
tf.summary.scalar('loss', loss_op, family='train'),
tf.summary.scalar('image_loss', image_loss_op, family='train'),
tf.summary.scalar('lpip_loss', lpips_loss_op, family='train'),
tf.summary.scalar('G_loss', G_loss, family='train'),
tf.summary.scalar('secret_loss', secret_loss_op, family='train'),
tf.summary.scalar('dis_loss', D_loss, family='train'),
tf.summary.scalar('Y_loss', yuv_loss_op[0], family='color_loss'),
tf.summary.scalar('U_loss', yuv_loss_op[1], family='color_loss'),
tf.summary.scalar('V_loss', yuv_loss_op[2], family='color_loss'),
] + transform_summaries)
image_summary_op = tf.summary.merge([
image_to_summary(image_input, 'image_input', family='input'),
image_to_summary(input_warped, 'image_warped', family='input'),
image_to_summary(encoded_warped, 'encoded_warped', family='encoded'),
image_to_summary(residual_warped+.5, 'residual', family='encoded'),
image_to_summary(encoded_image, 'encoded_image', family='encoded'),
image_to_summary(transformed_image, 'transformed_image', family='transformed'),
image_to_summary(D_heatmap, 'discriminator', family='losses'),
])
return loss_op, secret_loss_op, D_loss, summary_op, image_summary_op, bit_acc
def image_to_summary(image, name, family='train'):
image = tf.clip_by_value(image, 0, 1)
image = tf.cast(image * 255, dtype=tf.uint8)
summary = tf.summary.image(name,image,max_outputs=1,family=family)
return summary
def prepare_deployment_hiding_graph(encoder, secret_input, image_input):
residual = encoder((secret_input, image_input))
encoded_image = residual + image_input
encoded_image = tf.clip_by_value(encoded_image, 0, 1)
return encoded_image, residual
def prepare_deployment_reveal_graph(decoder, image_input):
decoded_secret = decoded_secret = decoder(image_input)
return tf.round(tf.sigmoid(decoded_secret))