forked from ynulonger/DE_CNN
-
Notifications
You must be signed in to change notification settings - Fork 1
/
get_1D_data.py
141 lines (118 loc) · 5.51 KB
/
get_1D_data.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
import os
import sys
import math
import numpy as np
import pandas as pd
import scipy.io as sio
from sklearn import preprocessing
from scipy.signal import butter, lfilter
def butter_bandpass(lowcut, highcut, fs, order=5):
nyq = 0.5 * fs
low = lowcut / nyq
high = highcut / nyq
b, a = butter(order, [low, high], btype='band')
return b, a
def butter_bandpass_filter(data, lowcut, highcut, fs, order=5):
b, a = butter_bandpass(lowcut, highcut, fs, order=order)
y = lfilter(b, a, data)
return y
def read_file(file):
data = sio.loadmat(file)
data = data['data']
# print(data.shape)
return data
def compute_DE(signal):
variance = np.var(signal,ddof=1)
return math.log(2*math.pi*math.e*variance)/2
def decompose(file):
# trial*channel*sample
start_index = 384 #3s pre-trial signals
data = read_file(file)
shape = data.shape
frequency = 128
decomposed_de = np.empty([0,4,60])
base_DE = np.empty([0,128])
for trial in range(40):
temp_base_DE = np.empty([0])
temp_base_theta_DE = np.empty([0])
temp_base_alpha_DE = np.empty([0])
temp_base_beta_DE = np.empty([0])
temp_base_gamma_DE = np.empty([0])
temp_de = np.empty([0,60])
for channel in range(32):
trial_signal = data[trial,channel,384:]
base_signal = data[trial,channel,:384]
#****************compute base DE****************
base_theta = butter_bandpass_filter(base_signal, 4, 8, frequency, order=3)
base_alpha = butter_bandpass_filter(base_signal, 8,14, frequency, order=3)
base_beta = butter_bandpass_filter(base_signal,14,31, frequency, order=3)
base_gamma = butter_bandpass_filter(base_signal,31,45, frequency, order=3)
base_theta_DE = (compute_DE(base_theta[:128])+compute_DE(base_theta[128:256])+compute_DE(base_theta[256:]))/3
base_alpha_DE =(compute_DE(base_alpha[:128])+compute_DE(base_alpha[128:256])+compute_DE(base_alpha[256:]))/3
base_beta_DE =(compute_DE(base_beta[:128])+compute_DE(base_beta[128:256])+compute_DE(base_beta[256:]))/3
base_gamma_DE =(compute_DE(base_gamma[:128])+compute_DE(base_gamma[128:256])+compute_DE(base_gamma[256:]))/3
temp_base_theta_DE = np.append(temp_base_theta_DE,base_theta_DE)
temp_base_gamma_DE = np.append(temp_base_gamma_DE,base_gamma_DE)
temp_base_beta_DE = np.append(temp_base_beta_DE,base_beta_DE)
temp_base_alpha_DE = np.append(temp_base_alpha_DE,base_alpha_DE)
theta = butter_bandpass_filter(trial_signal, 4, 8, frequency, order=3)
alpha = butter_bandpass_filter(trial_signal, 8, 14, frequency, order=3)
beta = butter_bandpass_filter(trial_signal, 14, 31, frequency, order=3)
gamma = butter_bandpass_filter(trial_signal, 31, 45, frequency, order=3)
DE_theta = np.zeros(shape=[0],dtype = float)
DE_alpha = np.zeros(shape=[0],dtype = float)
DE_beta = np.zeros(shape=[0],dtype = float)
DE_gamma = np.zeros(shape=[0],dtype = float)
for index in range(60):
DE_theta =np.append(DE_theta,compute_DE(theta[index*frequency:(index+1)*frequency]))
DE_alpha =np.append(DE_alpha,compute_DE(alpha[index*frequency:(index+1)*frequency]))
DE_beta =np.append(DE_beta,compute_DE(beta[index*frequency:(index+1)*frequency]))
DE_gamma =np.append(DE_gamma,compute_DE(gamma[index*frequency:(index+1)*frequency]))
temp_de = np.vstack([temp_de,DE_theta])
temp_de = np.vstack([temp_de,DE_alpha])
temp_de = np.vstack([temp_de,DE_beta])
temp_de = np.vstack([temp_de,DE_gamma])
temp_trial_de = temp_de.reshape(-1,4,60)
decomposed_de = np.vstack([decomposed_de,temp_trial_de])
temp_base_DE = np.append(temp_base_theta_DE,temp_base_alpha_DE)
temp_base_DE = np.append(temp_base_DE,temp_base_beta_DE)
temp_base_DE = np.append(temp_base_DE,temp_base_gamma_DE)
base_DE = np.vstack([base_DE,temp_base_DE])
decomposed_de = decomposed_de.reshape(-1,32,4,60).transpose([0,3,2,1]).reshape(-1,4,32).reshape(-1,128)
print("base_DE shape:",base_DE.shape)
print("trial_DE shape:",decomposed_de.shape)
return base_DE,decomposed_de
def get_labels(file):
#0 valence, 1 arousal, 2 dominance, 3 liking
valence_labels = sio.loadmat(file)["labels"][:,0]>5 # valence labels
arousal_labels = sio.loadmat(file)["labels"][:,1]>5 # arousal labels
final_valence_labels = np.empty([0])
final_arousal_labels = np.empty([0])
for i in range(len(valence_labels)):
for j in range(0,60):
final_valence_labels = np.append(final_valence_labels,valence_labels[i])
final_arousal_labels = np.append(final_arousal_labels,arousal_labels[i])
print("labels:",final_arousal_labels.shape)
return final_arousal_labels,final_valence_labels
def wgn(x, snr):
snr = 10**(snr/10.0)
xpower = np.sum(x**2)/len(x)
npower = xpower / snr
return np.random.randn(len(x)) * np.sqrt(npower)
def feature_normalize(data):
mean = data[data.nonzero()].mean()
sigma = data[data. nonzero ()].std()
data_normalized = data
data_normalized[data_normalized.nonzero()] = (data_normalized[data_normalized.nonzero()] - mean)/sigma
return data_normalized
if __name__ == '__main__':
dataset_dir = "/home/data_preprocessed_matlab/"
result_dir = "/home/yyl/DE_CNN/1D_dataset/"
if os.path.isdir(result_dir)==False:
os.makedirs(result_dir)
for file in os.listdir(dataset_dir):
print("processing: ",file,"......")
file_path = os.path.join(dataset_dir,file)
base_DE,trial_DE = decompose(file_path)
arousal_labels,valence_labels = get_labels(file_path)
sio.savemat(result_dir+"DE_"+file,{"base_data":base_DE,"data":trial_DE,"valence_labels":valence_labels,"arousal_labels":arousal_labels})