-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathconcatenating_features.py
158 lines (112 loc) · 5.97 KB
/
concatenating_features.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
import tensorflow as tf
import numpy as np
import util
# concatenating envelope and mel spectrogram features together
def lstm_env_mel_spatial_filter(shape_eeg, shape_feature1, shape_feature2, units_lstm=32, filters_cnn_eeg=16, filters_cnn_env=16,
units_hidden=128,
stride_temporal=3, kerSize_temporal=9, spatial_filters_eeg=32,
spatial_filters_mel=8, fun_act='tanh'):
############
input_eeg = tf.keras.layers.Input(shape=shape_eeg)
env1 = tf.keras.layers.Input(shape=shape_feature1)
env2 = tf.keras.layers.Input(shape=shape_feature1)
mel1 = tf.keras.layers.Input(shape=shape_feature2)
mel2 = tf.keras.layers.Input(shape=shape_feature2)
############
#### upper part of network dealing with EEG.
layer_exp1 = tf.keras.layers.Lambda(lambda x: tf.keras.backend.expand_dims(x, axis=3))
eeg_proj = input_eeg
# layer
output_eeg = tf.keras.layers.BatchNormalization()(eeg_proj) # batch normalization
output_eeg = tf.keras.layers.Conv1D(spatial_filters_eeg, kernel_size=1)(output_eeg)
# layer
output_eeg = tf.keras.layers.BatchNormalization()(output_eeg)
output_eeg = layer_exp1(output_eeg)
output_eeg = tf.keras.layers.Convolution2D(filters_cnn_eeg, (kerSize_temporal, 1),
strides=(stride_temporal, 1), activation="relu")(output_eeg)
# layer
layer_permute = tf.keras.layers.Permute((1, 3, 2))
output_eeg = layer_permute(output_eeg)
layer_reshape = tf.keras.layers.Reshape((tf.keras.backend.int_shape(output_eeg)[1],
tf.keras.backend.int_shape(output_eeg)[2] *
tf.keras.backend.int_shape(output_eeg)[3]))
output_eeg = layer_reshape(output_eeg)
layer2_timeDis = tf.keras.layers.TimeDistributed(tf.keras.layers.Dense(units_hidden, activation=fun_act))
output_eeg = layer2_timeDis(output_eeg)
# layer
output_eeg = tf.keras.layers.BatchNormalization()(output_eeg)
layer3_timeDis = tf.keras.layers.TimeDistributed(tf.keras.layers.Dense(units_lstm, activation=fun_act))
output_eeg = layer3_timeDis(output_eeg)
##############
#### Bottom part of the network dealing with Speech.
env1_proj = env1
env2_proj = env2
# layer
BN_layer = tf.keras.layers.BatchNormalization()
output_env1 = BN_layer(env1_proj)
output_env2 = BN_layer(env2_proj)
output_env1 = layer_exp1(output_env1)
output_env2 = layer_exp1(output_env2)
conv_env_layer = tf.keras.layers.Convolution2D(filters_cnn_env, (kerSize_temporal, 1),
strides=(stride_temporal, 1), activation="relu")
output_env1 = conv_env_layer(output_env1)
output_env2 = conv_env_layer(output_env2)
## speech feature 2
mel1_proj = mel1
mel2_proj = mel2
# layer
BN_layer = tf.keras.layers.BatchNormalization()
output_mel1 = BN_layer(mel1_proj)
output_mel2 = BN_layer(mel2_proj)
env_spatial_layer = tf.keras.layers.Conv1D(spatial_filters_mel, kernel_size=1)
output_mel1 = env_spatial_layer(output_mel1)
output_mel2 = env_spatial_layer(output_mel2)
# layer
BN_layer1 = tf.keras.layers.BatchNormalization()
output_mel1 = BN_layer1(output_mel1)
output_mel2 = BN_layer1(output_mel2)
output_mel1 = layer_exp1(output_mel1)
output_mel2 = layer_exp1(output_mel2)
conv_env_layer = tf.keras.layers.Convolution2D(filters_cnn_env, (kerSize_temporal, 1),
strides=(stride_temporal, 1), activation="relu")
output_mel1 = conv_env_layer(output_mel1)
output_mel2 = conv_env_layer(output_mel2)
# layer: combine two features
layer_permute = tf.keras.layers.Permute((1, 3, 2))
output_env1 = layer_permute(output_env1)
output_env2 = layer_permute(output_env2)
layer_reshape = tf.keras.layers.Reshape((tf.keras.backend.int_shape(output_env1)[1],
tf.keras.backend.int_shape(output_env1)[2] *
tf.keras.backend.int_shape(output_env1)[3]))
output_env1 = layer_reshape(output_env1) # size = (210,32)
output_env2 = layer_reshape(output_env2)
output_mel1 = layer_permute(output_mel1)
output_mel2 = layer_permute(output_mel2)
layer_reshape = tf.keras.layers.Reshape((tf.keras.backend.int_shape(output_mel1)[1],
tf.keras.backend.int_shape(output_mel1)[2] *
tf.keras.backend.int_shape(output_mel1)[3]))
output_mel1 = layer_reshape(output_mel1)
output_mel2 = layer_reshape(output_mel2)
output_spch1 = tf.keras.layers.Concatenate()([output_env1, output_mel1])
output_spch2 = tf.keras.layers.Concatenate()([output_env2, output_mel2])
# lstm_spch = tf.keras.layers.LSTM(units_lstm, return_sequences=True, activation= fun_act)
lstm_spch = tf.keras.layers.CuDNNLSTM(units_lstm, return_sequences=True)
output_spch1 = lstm_spch(output_spch1)
output_spch2 = lstm_spch(output_spch2)
##############
#### last common layers
# layer
layer_dot = util.DotLayer()
cos_scores = layer_dot([output_eeg, output_spch1])
cos_scores2 = layer_dot([output_eeg, output_spch2])
# layer
layer_expand = tf.keras.layers.Lambda(lambda x: tf.keras.backend.expand_dims(x, axis=2))
layer_sigmoid = tf.keras.layers.TimeDistributed(tf.keras.layers.Dense(1, activation='sigmoid'))
cos_scores_mix = tf.keras.layers.Concatenate()([layer_expand(cos_scores), layer_expand(cos_scores2)])
cos_scores_sig = layer_sigmoid(cos_scores_mix)
# layer
layer_ave = tf.keras.layers.Lambda(lambda x: tf.reduce_mean(x, axis=1, keepdims=True))
cos_scores_sig = util.SqueezeLayer()(cos_scores_sig, axis=2)
y_out = layer_ave(cos_scores_sig)
model = tf.keras.Model(inputs=[input_eeg, env1, env2, mel1, mel2], outputs=[y_out, cos_scores_sig])
return model