-
Notifications
You must be signed in to change notification settings - Fork 2
/
vision_module.py
373 lines (295 loc) · 16.6 KB
/
vision_module.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
import tensorflow as tf
l2_reg = tf.keras.regularizers.l2(l=0.1)
class TransporterEncoder(tf.keras.Model):
def __init__(self):
super(TransporterEncoder, self).__init__()
self.conv_1 = tf.keras.layers.Conv2D(filters=16, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_1 = tf.keras.layers.BatchNormalization()
self.relu_1 = tf.keras.layers.ReLU()
self.conv_2 = tf.keras.layers.Conv2D(filters=16, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_2 = tf.keras.layers.BatchNormalization()
self.relu_2 = tf.keras.layers.ReLU()
self.conv_3 = tf.keras.layers.Conv2D(filters=32, kernel_size=3, strides=2,
padding="same", kernel_regularizer=l2_reg)
self.bn_3 = tf.keras.layers.BatchNormalization()
self.relu_3 = tf.keras.layers.ReLU()
self.conv_4 = tf.keras.layers.Conv2D(filters=32, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_4 = tf.keras.layers.BatchNormalization()
self.relu_4 = tf.keras.layers.ReLU()
self.filters = [16, 16, 32, 32]
self.kernels = [3, 3, 3, 3]
self.strides = [1, 1, 2, 1]
def call(self, inputs, training=True):
h1 = self.relu_1(self.bn_1(self.conv_1(inputs), training=training))
h2 = self.relu_2(self.bn_2(self.conv_2(h1), training=training))
h3 = self.relu_3(self.bn_3(self.conv_3(h2), training=training))
h4 = self.relu_4(self.bn_4(self.conv_4(h3), training=training))
return h4
class TransporterDecoder(tf.keras.Model):
def __init__(self, colour):
super(TransporterDecoder, self).__init__()
self.deconv_1 = tf.keras.layers.Conv2D(filters=128, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_1 = tf.keras.layers.BatchNormalization()
self.relu_1 = tf.keras.layers.ReLU()
self.deconv_2 = tf.keras.layers.Conv2D(filters=128, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_2 = tf.keras.layers.BatchNormalization()
self.relu_2 = tf.keras.layers.ReLU()
self.deconv_3 = tf.keras.layers.Conv2D(filters=64, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_3 = tf.keras.layers.BatchNormalization()
self.relu_3 = tf.keras.layers.ReLU()
self.deconv_4 = tf.keras.layers.Conv2D(filters=32, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_4 = tf.keras.layers.BatchNormalization()
self.relu_4 = tf.keras.layers.ReLU()
if colour:
self.deconv_5 = tf.keras.layers.Conv2D(filters=3, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_5 = tf.keras.layers.BatchNormalization()
elif not colour:
self.deconv_5 = tf.keras.layers.Conv2D(filters=1, kernel_size=3, strides=1,
padding="same", kernel_regularizer=l2_reg)
self.bn_5 = tf.keras.layers.BatchNormalization()
def call(self, features, training=True):
h1 = self.relu_1(self.bn_1(self.deconv_1(features), training=training))
h2 = self.relu_2(self.bn_2(self.deconv_2(h1), training=training))
# 2X upsample
h2_shape = h2.shape
h2 = tf.image.resize(h2, (2*h2_shape[1], 2*h2_shape[2]), method='bilinear')
h3 = self.relu_3(self.bn_3(self.deconv_3(h2), training=training))
h4 = self.relu_4(self.bn_4(self.deconv_4(h3), training=training))
reconstructed_image = tf.keras.activations.sigmoid(self.bn_5(self.deconv_5(h4), training=training))
return reconstructed_image
def get_keypoint_data_from_feature_map(feature_map, gauss_std):
"""Returns keypoint information from a feature map.
Args:
feature_map: [B, H, W, K] Tensor, should be activations from a convnet.
gauss_std: float, the standard deviation of the gaussians to be put around
the keypoints.
Returns:
a dict with keys:
'centers': A tensor of shape [B, K, 2] of the center locations for each
of the K keypoints.
'heatmaps': A tensor of shape [B, H, W, K] of gaussian maps over the
keypoints.
"""
gauss_mu = _get_keypoint_mus(feature_map)
map_size = feature_map.shape.as_list()[1:3]
gauss_maps = _get_gaussian_maps(gauss_mu, map_size, 1.0 / gauss_std)
return gauss_mu, gauss_maps
def _get_keypoint_mus(keypoint_features):
"""Returns the keypoint center points.
Args:
keypoint_features: A tensor of shape [B, F_h, F_w, K] where K is the number
of keypoints to extract.
Returns:
A tensor of shape [B, K, 2] of the y, x center points of each keypoint. Each
center point are in the range [-1, 1]^2. Note: the first element is the y
coordinate, the second is the x coordinate.
"""
gauss_y = _get_coord(keypoint_features, 1)
gauss_x = _get_coord(keypoint_features, 2)
gauss_mu = tf.stack([gauss_y, gauss_x], axis=2)
return gauss_mu
def _get_coord(features, axis):
"""Returns the keypoint coordinate encoding for the given axis.
Args:
features: A tensor of shape [B, F_h, F_w, K] where K is the number of
keypoints to extract.
axis: `int` which axis to extract the coordinate for. Has to be axis 1 or 2.
Returns:
A tensor of shape [B, K] containing the keypoint centers along the given
axis. The location is given in the range [-1, 1].
"""
if axis != 1 and axis != 2:
raise ValueError("Axis needs to be 1 or 2.")
other_axis = 1 if axis == 2 else 2
axis_size = features.shape[axis]
# Compute the normalized weight for each row/column along the axis
g_c_prob = tf.reduce_mean(features, axis=other_axis)
g_c_prob = tf.nn.softmax(g_c_prob, axis=1)
# Linear combination of the interval [-1, 1] using the normalized weights to
# give a single coordinate in the same interval [-1, 1]
scale = tf.cast(tf.linspace(-1.0, 1.0, axis_size), tf.float32)
scale = tf.reshape(scale, [1, axis_size, 1])
coordinate = tf.reduce_sum(g_c_prob * scale, axis=1)
return coordinate
def _get_gaussian_maps(mu, map_size, inv_std, power=2):
"""Transforms the keypoint center points to a gaussian masks."""
mu_y, mu_x = mu[:, :, 0:1], mu[:, :, 1:2]
y = tf.cast(tf.linspace(-1.0, 1.0, map_size[0]), tf.float32)
x = tf.cast(tf.linspace(-1.0, 1.0, map_size[1]), tf.float32)
mu_y, mu_x = tf.expand_dims(mu_y, -1), tf.expand_dims(mu_x, -1)
y = tf.reshape(y, [1, 1, map_size[0], 1])
x = tf.reshape(x, [1, 1, 1, map_size[1]])
g_y = tf.pow(y - mu_y, power)
g_x = tf.pow(x - mu_x, power)
dist = (g_y + g_x) * tf.pow(inv_std, power)
g_yx = tf.exp(-dist)
g_yx = tf.transpose(g_yx, perm=[0, 2, 3, 1])
return g_yx
class TransporterKeypointer(tf.keras.Model):
"""Module for extracting keypoints from an image."""
def __init__(self, num_keypoints, gauss_std):
super(TransporterKeypointer, self).__init__()
self.num_keypoints = num_keypoints
self.gauss_std = gauss_std
self.keypoint_encoder = TransporterEncoder()
self.conv = tf.keras.layers.Conv2D(filters=self.num_keypoints, kernel_size=1,
strides=1, kernel_regularizer=l2_reg)
def call(self, image, training=True):
image_features = self.keypoint_encoder(image, training=training)
keypoint_features = self.conv(image_features)
return get_keypoint_data_from_feature_map(keypoint_features, self.gauss_std)
class ConvEncoder(tf.keras.Model):
def __init__(self, latent_dim_size):
super(ConvEncoder, self).__init__()
self.conv_1 = tf.keras.layers.Conv2D(filters=32, kernel_size=4, strides=1,
padding="same")
self.bn_1 = tf.keras.layers.BatchNormalization()
self.relu_1 = tf.keras.layers.ReLU()
self.conv_2 = tf.keras.layers.Conv2D(filters=64, kernel_size=3, strides=2,
padding="same")
self.bn_2 = tf.keras.layers.BatchNormalization()
self.relu_2 = tf.keras.layers.ReLU()
self.conv_3 = tf.keras.layers.Conv2D(filters=64, kernel_size=3, strides=2,
padding="same")
self.bn_3 = tf.keras.layers.BatchNormalization()
self.relu_3 = tf.keras.layers.ReLU()
self.conv_4 = tf.keras.layers.Conv2D(filters=128, kernel_size=3, strides=1,
padding="same")
self.bn_4 = tf.keras.layers.BatchNormalization()
self.relu_4 = tf.keras.layers.ReLU()
self.flatten = tf.keras.layers.Flatten()
self.fc = tf.keras.layers.Dense(2*latent_dim_size) # no activation fc_size = size(mu) + size(sigma)
# useful vars
self.filters = [32, 64, 64, 128]
self.kernels = [4, 3, 3, 3]
self.strides = [1, 2, 2, 1]
def call(self, inputs, training=True):
h1 = self.relu_1(self.bn_1(self.conv_1(inputs), training=training))
h2 = self.relu_2(self.bn_2(self.conv_2(h1), training=training))
h3 = self.relu_3(self.bn_3(self.conv_3(h2), training=training))
h4 = self.relu_4(self.bn_4(self.conv_4(h3), training=training))
flatten_h4 = self.flatten(h4)
mean, var = tf.split(self.fc(flatten_h4), num_or_size_splits=2, axis=1)
return mean, tf.nn.softplus(var), [h1, h2, h3, h4] # returning +ve vars
class DeconvDecoder(tf.keras.Model):
def __init__(self, colour):
super(DeconvDecoder, self).__init__()
self.dense_1 = tf.keras.layers.Dense(15*15)
self.relu_d = tf.keras.layers.ReLU()
self.deconv_1 = tf.keras.layers.Conv2DTranspose(filters=128, kernel_size=4, strides=1)
self.bn_1 = tf.keras.layers.BatchNormalization()
self.relu_1 = tf.keras.layers.ReLU()
self.deconv_2 = tf.keras.layers.Conv2DTranspose(filters=64, kernel_size=4, strides=2)
self.bn_2 = tf.keras.layers.BatchNormalization()
self.relu_2 = tf.keras.layers.ReLU()
self.deconv_3 = tf.keras.layers.Conv2DTranspose(filters=64, kernel_size=4, strides=2)
self.bn_3 = tf.keras.layers.BatchNormalization()
self.relu_3 = tf.keras.layers.ReLU()
self.deconv_4 = tf.keras.layers.Conv2DTranspose(filters=32, kernel_size=4, strides=1)
self.bn_4 = tf.keras.layers.BatchNormalization()
self.relu_4 = tf.keras.layers.ReLU()
if colour: # color image
self.deconv_5 = tf.keras.layers.Conv2DTranspose(filters=3, kernel_size=4, strides=1)
self.bn_5 = tf.keras.layers.BatchNormalization()
elif not colour: # greyscale image
self.deconv_5 = tf.keras.layers.Conv2DTranspose(filters=1, kernel_size=4, strides=1)
self.bn_5 = tf.keras.layers.BatchNormalization()
def call(self, z, training=True):
# flattened latent z need to reshape to [batch, H, W, C]
dense_z = self.relu_d(self.dense_1(z))
z_3d = tf.reshape(dense_z, [-1, 15, 15, 1])
h1 = self.relu_1(self.bn_1(self.deconv_1(z_3d), training=training))
h2 = self.relu_2(self.bn_2(self.deconv_2(h1), training=training))
h3 = self.relu_3(self.bn_3(self.deconv_3(h2), training=training))
h4 = self.relu_4(self.bn_4(self.deconv_4(h3), training=training))
h5 = tf.keras.activations.sigmoid(self.bn_5(self.deconv_5(h4), training=training)) # img re-scaled to vals [0-1] -> sigmoid output used for decoder last layer
return h5
class ConvDecoder(tf.keras.Model):
def __init__(self, colour, img_size):
super(ConvDecoder, self).__init__()
self.img_size = img_size
self.dense_1 = tf.keras.layers.Dense(int(self.img_size/4)*int(self.img_size/4))
self.relu_d = tf.keras.layers.ReLU()
self.conv_1 = tf.keras.layers.Conv2D(filters=128, kernel_size=3, strides=1,
padding="same")
self.bn_1 = tf.keras.layers.BatchNormalization()
self.relu_1 = tf.keras.layers.ReLU()
self.conv_2 = tf.keras.layers.Conv2D(filters=64, kernel_size=3, strides=1,
padding="same")
self.bn_2 = tf.keras.layers.BatchNormalization()
self.relu_2 = tf.keras.layers.ReLU()
self.conv_3 = tf.keras.layers.Conv2D(filters=64, kernel_size=3, strides=1,
padding="same")
self.bn_3 = tf.keras.layers.BatchNormalization()
self.relu_3 = tf.keras.layers.ReLU()
self.conv_4 = tf.keras.layers.Conv2D(filters=32, kernel_size=3, strides=1,
padding="same")
self.bn_4 = tf.keras.layers.BatchNormalization()
self.relu_4 = tf.keras.layers.ReLU()
self.conv_5 = tf.keras.layers.Conv2D(filters=16, kernel_size=4, strides=1,
padding="same")
self.bn_5 = tf.keras.layers.BatchNormalization()
self.relu_5 = tf.keras.layers.ReLU()
if colour: # color image
self.conv_6 = tf.keras.layers.Conv2D(filters=3, kernel_size=4, strides=1,
padding="same")
self.bn_6 = tf.keras.layers.BatchNormalization()
elif not colour: # greyscale image
self.conv_6 = tf.keras.layers.Conv2D(filters=1, kernel_size=4, strides=1,
padding="same")
self.bn_6 = tf.keras.layers.BatchNormalization()
def call(self, z, training=True):
# flattened latent z need to reshape to [batch, H, W, C]
dense_z = self.relu_d(self.dense_1(z))
z_3d = tf.reshape(dense_z, [-1, int(self.img_size/4), int(self.img_size/4), 1])
h1 = self.relu_1(self.bn_1(self.conv_1(z_3d), training=training)) # (21, 21)
# upsample 2x
h1_shape = h1.shape
h1 = tf.image.resize(h1, (2*h1_shape[1], 2*h1_shape[2]), method='nearest')
h2 = self.relu_2(self.bn_2(self.conv_2(h1), training=training)) # (42, 42)
# upsample 2x
h2_shape = h2.shape
h2 = tf.image.resize(h2, (2*h2_shape[1], 2*h2_shape[2]), method='nearest') # (84, 84)
h3 = self.relu_3(self.bn_3(self.conv_3(h2), training=training))
h4 = self.relu_4(self.bn_4(self.conv_4(h3), training=training))
h5 = self.relu_5(self.bn_5(self.conv_5(h4), training=training))
h6 = tf.keras.activations.sigmoid(self.bn_6(self.conv_6(h5), training=training)) # img re-scaled to vals [0-1] -> sigmoid output used for decoder last layer
return h6
class Pnet(tf.keras.Model):
def __init__(self, num_keypoints, gauss_std):
super(Pnet, self).__init__()
self.num_keypoints = num_keypoints
self.gauss_std = gauss_std
self.conv_1 = tf.keras.layers.Conv2D(filters=32, kernel_size=4,
strides=1, padding="same")
self.bn_1 = tf.keras.layers.BatchNormalization()
self.relu_1 = tf.keras.layers.ReLU()
self.conv_2 = tf.keras.layers.Conv2D(filters=64, kernel_size=3,
strides=2, padding="same")
self.bn_2 = tf.keras.layers.BatchNormalization()
self.relu_2 = tf.keras.layers.ReLU()
self.conv_3 = tf.keras.layers.Conv2D(filters=64, kernel_size=3,
strides=2, padding="same")
self.bn_3 = tf.keras.layers.BatchNormalization()
self.relu_3 = tf.keras.layers.ReLU()
self.conv_4 = tf.keras.layers.Conv2D(filters=128, kernel_size=3,
strides=1, padding="same")
self.bn_4 = tf.keras.layers.BatchNormalization()
self.relu_4 = tf.keras.layers.ReLU()
self.conv_5 = tf.keras.layers.Conv2D(filters=self.num_keypoints,
kernel_size=1, strides=1, padding="same")
def call(self, inputs, training=True):
h1 = self.relu_1(self.bn_1(self.conv_1(inputs), training=training))
h2 = self.relu_2(self.bn_2(self.conv_2(h1), training=training))
h3 = self.relu_3(self.bn_3(self.conv_3(h2), training=training))
h4 = self.relu_4(self.bn_4(self.conv_4(h3), training=training))
keypoint_features = self.conv_5(h4)
return get_keypoint_data_from_feature_map(keypoint_features, self.gauss_std)