-
Notifications
You must be signed in to change notification settings - Fork 5
/
Augment.py
163 lines (137 loc) · 5.64 KB
/
Augment.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
import tensorflow as tf
import numpy as np
import random
from scipy.interpolate import interp1d
#Part of the code reference https://github.com/iantangc/ContrastiveLearningHAR
def resampling_fast(x,M,N):
time_steps = x.shape[1]
raw_set = np.arange(time_steps)
interp_steps = np.arange(0, raw_set[-1] + 1e-1, 1 / (M + 1))
x_interp = interp1d(raw_set, x, axis=1)
x_up = x_interp(interp_steps)
length_inserted = x_up.shape[1]
start = random.randint(0, length_inserted - time_steps * (N + 1))
index_selected = np.arange(start, start + time_steps * (N + 1), N + 1)
return x_up[:, index_selected, :]
def resampling_fast_random(x):
M, N = random.choice([[1, 0], [2, 1], [3, 2]])
time_steps = x.shape[1]
raw_set = np.arange(x.shape[1])
interp_steps = np.arange(0, raw_set[-1] + 1e-1, 1 / (M + 1))
x_interp = interp1d(raw_set, x, axis=1)
x_up = x_interp(interp_steps)
length_inserted = x_up.shape[1]
start = random.randint(0, length_inserted - time_steps * (N + 1))
index_selected = np.arange(start, start + time_steps * (N + 1), N + 1)
return x_up[:, index_selected, :]
def resampling(x,M,N):
'''
:param x: the data of a batch,shape=(batch_size,timesteps,features)
:param M: the number of new value under tow values
:param N: the interval of resampling
:return: x after resampling，shape=(batch_size,timesteps,features)
'''
assert M>N,'the value of M have to greater than N'
timesetps = x.shape[1]
for i in range(timesetps-1):
x1 = x[:,i*(M+1),:]
x2 = x[:,i*(M+1)+1,:]
for j in range(M):
v = np.add(x1,np.subtract(x2,x1)*(j+1)/(M+1))
x = np.insert(x,i*(M+1)+j+1,v,axis=1)
length_inserted = x.shape[1]
start = random.randint(0,length_inserted-timesetps*(N+1))
index_selected = np.arange(start,start+timesetps*(N+1),N+1)
return x[:,index_selected,:]
return x
def resampling_random(x):
import random
M = random.randint(1, 3)
N = random.randint(0, M - 1)
assert M > N, 'the value of M have to greater than N'
timesetps = x.shape[1]
for i in range(timesetps - 1):
x1 = x[:, i * (M + 1), :]
x2 = x[:, i * (M + 1) + 1, :]
for j in range(M):
v = np.add(x1, np.subtract(x2, x1) * (j + 1) / (M + 1))
x = np.insert(x, i * (M + 1) + j + 1, v, axis=1)
length_inserted = x.shape[1]
num = x.shape[0]
start = random.randint(0, length_inserted - timesetps * (N + 1))
index_selected = np.arange(start, start + timesetps * (N + 1), N + 1)
x_selected=x[0,index_selected,:][np.newaxis,]
for k in range(1,num):
start = random.randint(0, length_inserted - timesetps * (N + 1))
index_selected = np.arange(start, start + timesetps * (N + 1), N + 1)
x_selected = np.concatenate((x_selected,x[k,index_selected,:][np.newaxis,]),axis=0)
return x_selected
def noise(x):
x = tf.add(x,tf.multiply(x,tf.cast(tf.random.uniform(shape = (x.shape[0],x.shape[1],x.shape[2]),minval=-0.1,maxval=0.1),tf.float64)))
return x
def rotate(x,angles=np.pi/12):
t = angles
f = angles
r = angles
Rx = np.array([[1, 0, 0],
[0, np.cos(t), -np.sin(t)],
[0, np.sin(t), np.cos(t)]])
Ry = np.array([[np.cos(f), 0, np.sin(f)],
[0, 1, 0],
[-np.sin(f), 1, np.cos(f)]])
Rz = np.array([[np.cos(r), -np.sin(r), 0],
[np.sin(r), np.cos(r), 0],
[0, 0, 1]])
c = x.shape[2]//3
x_new = np.matmul(np.matmul(np.matmul(Rx,Ry),Rz),np.transpose(x[:,:,0:3],(0,2,1))).transpose(0,2,1)
for i in range(1,c):
temp = np.matmul(np.matmul(np.matmul(Rx,Ry),Rz),np.transpose(x[:,:,i*3:i*3+3],(0,2,1))).transpose(0,2,1)
x_new = np.concatenate((x_new,temp),axis=-1)
return x_new
def scaling(x):
alpha = np.random.randint(7,10)/10
# alpha = 0.9
return tf.multiply(x,alpha)
#
def magnify(x):
lam = np.random.randint(11,14)/10
return tf.multiply(x,lam)
def inverting(x):
return np.multiply(x,-1)
def reversing(x):
return x[:,-1::-1,:]
def rotation(x):
c = x.shape[2]//3
x_new = rotation_transform_vectorized(x[:,:,0:3])
for i in range(1,c):
temp = rotation_transform_vectorized(x[:,:,i*3:(i+1)*3])
x_new = np.concatenate((x_new,temp),axis=-1)
return x_new
def rotation_transform_vectorized(X):
"""
Applying a random 3D rotation
"""
axes = np.random.uniform(low=-1, high=1, size=(X.shape[0], X.shape[2]))
angles = np.random.uniform(low=-np.pi, high=np.pi, size=(X.shape[0]))
matrices = axis_angle_to_rotation_matrix_3d_vectorized(axes, angles)
return np.matmul(X, matrices)
def axis_angle_to_rotation_matrix_3d_vectorized(axes, angles):
"""
Get the rotational matrix corresponding to a rotation of (angle) radian around the axes
Reference: the Transforms3d package - transforms3d.axangles.axangle2mat
Formula: http://en.wikipedia.org/wiki/Rotation_matrix#Axis_and_angle
"""
axes = axes / np.linalg.norm(axes, ord=2, axis=1, keepdims=True)
x = axes[:, 0]; y = axes[:, 1]; z = axes[:, 2]
c = np.cos(angles)
s = np.sin(angles)
C = 1 - c
xs = x*s; ys = y*s; zs = z*s
xC = x*C; yC = y*C; zC = z*C
xyC = x*yC; yzC = y*zC; zxC = z*xC
m = np.array([
[ x*xC+c, xyC-zs, zxC+ys ],
[ xyC+zs, y*yC+c, yzC-xs ],
[ zxC-ys, yzC+xs, z*zC+c ]])
matrix_transposed = np.transpose(m, axes=(2,0,1))
return matrix_transposed