-
Notifications
You must be signed in to change notification settings - Fork 0
/
load_data.py
187 lines (163 loc) · 6.58 KB
/
load_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
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
import pickle
import numpy as np
import math
import random
import os
def load_adni(r_path, cats=('AD', 'CN'), lbl=(1, 0)):
sids = []
mris = []
stas = []
labels = []
for g in ['ADNI1', 'ADNI2', 'ADNI3']:
for c in cats:
f_path = f'{r_path}/{g}-{c}.npy'
if not os.path.exists(f_path):
continue # skip when file is not exists
with open(f_path, 'rb') as f:
data = pickle.load(f) # (mri,[])
sids.extend(list(data.keys()))
data = list(data.values())
for mri, sta in data:
mris.append(mri)
stas.append(sta)
label = [lbl[cats.index(c)] for _ in data]
labels.extend(label)
print('{}-{} contains {} mris'.format(g, c, len(data)))
print('Dataset length: ', len(sids), len(mris), len(stas), len(labels))
# fix the random seed
np.random.seed(2022)
index = np.random.permutation(np.arange(len(mris)))
mris, labels, stas, sids = np.asarray(mris), np.asarray(labels), np.asarray(stas), np.asarray(sids)
mris, labels, stas, sids = mris[index], labels[index], stas[index], sids[index]
return mris, labels, stas, sids
def load_aibl(r_path, cats=('AD', 'CN'), lbl=(1, 0)):
sids = []
mris = []
stas = []
labels = []
g = 'AIBL'
for c in cats:
f_path = f'{r_path}/{g}-{c}.npy'
with open(f_path, 'rb') as f:
data = pickle.load(f) # (mri,[])
sids.extend(list(data.keys()))
data = list(data.values())
for mri, sta in data:
mris.append(mri)
stas.append(sta)
label = [lbl[cats.index(c)] for _ in data]
labels.extend(label)
print('{}-{} contains {} mris'.format(g, c, len(data)))
print('Dataset length: ', len(sids), len(mris), len(stas), len(labels))
mris, labels, stas, sids = np.asarray(mris), np.asarray(labels), np.asarray(stas), np.asarray(sids)
return mris, labels, stas, sids
def load_adni_balance(r_path, cats=('AD', 'CN'), lbl=(1, 0), rate=1.1):
"""
Sample the data by groups.
:param r_path: base path of the data
:param cats: groups you want to pick
:param lbl: group labels
:param rate: sample rate
:return: mris, labels, stas, sids
"""
sids_d = {}
mris_d = {}
stas_d = {}
labels_d = {}
for g in ['ADNI1', 'ADNI2', 'ADNI3']:
for c in cats:
f_path = f'{r_path}/{g}-{c}.npy'
if not os.path.exists(f_path):
continue
if c not in sids_d.keys():
sids_d[c] = []
mris_d[c] = []
stas_d[c] = []
labels_d[c] = []
with open(f_path, 'rb') as f:
data = pickle.load(f) # (mri,[])
sids_d[c].extend(list(data.keys()))
data = list(data.values())
for mri, sta in data:
mris_d[c].append(mri)
stas_d[c].append(sta)
label = [lbl[cats.index(c)] for _ in data]
labels_d[c].extend(label)
print('{}-{} contains {} mris'.format(g, c, len(data)))
# fix the random seed
np.random.seed(2022)
random.seed(2022)
cat_nums = [len(sids_d[c]) for c in cats]
mn = min(cat_nums)
sm_len = math.floor(mn * rate)
for c in cats:
if len(sids_d[c]) / mn > rate:
sids_d[c] = random.sample(sids_d[c], sm_len)
mris_d[c] = random.sample(mris_d[c], sm_len)
stas_d[c] = random.sample(stas_d[c], sm_len)
labels_d[c] = random.sample(labels_d[c], sm_len)
mris, labels, stas, sids = [], [], [], []
for c in cats:
mris.extend(mris_d[c])
labels.extend(labels_d[c])
stas.extend(stas_d[c])
sids.extend(sids_d[c])
print('Dataset length: ', len(sids), len(mris), len(stas), len(labels))
# shuffle the dataset
index = np.random.permutation(np.arange(len(mris)))
mris, labels, stas, sids = np.asarray(mris), np.asarray(labels), np.asarray(stas), np.asarray(sids)
mris, labels, stas, sids = mris[index], labels[index], stas[index], sids[index]
return mris, labels, stas, sids
def load_adni_pmci_balance(r_path, pmci_ids, smci_ids, cats=('PMCI', 'SMCI'), lbl=(1, 0), rate=1.1):
sids_d = {}
mris_d = {}
stas_d = {}
labels_d = {}
for g in ['ADNI1', 'ADNI2', 'ADNI3']:
for c in cats:
f_path = f'{r_path}/{g}-{c}.npy'
if not os.path.exists(f_path):
continue
if c not in sids_d.keys():
sids_d[c] = []
mris_d[c] = []
stas_d[c] = []
labels_d[c] = []
with open(f_path, 'rb') as f:
data = pickle.load(f) # (mri,[])
if c == 'PMCI':
data = {key: data[key] for key in data if key in pmci_ids}
else:
data = {key: data[key] for key in data if key in smci_ids}
sids_d[c].extend(list(data.keys()))
data = list(data.values())
for mri, sta in data:
mris_d[c].append(mri)
stas_d[c].append(sta)
label = [lbl[cats.index(c)] for _ in data]
labels_d[c].extend(label)
print('{}-{} contains {} mris'.format(g, c, len(data)))
# fix the random seed
np.random.seed(2022)
random.seed(2022)
cat_nums = [len(sids_d[c]) for c in cats]
mn = min(cat_nums)
sm_len = math.floor(mn * rate)
for c in cats:
if len(sids_d[c]) / mn > rate:
sids_d[c] = random.sample(sids_d[c], sm_len)
mris_d[c] = random.sample(mris_d[c], sm_len)
stas_d[c] = random.sample(stas_d[c], sm_len)
labels_d[c] = random.sample(labels_d[c], sm_len)
mris, labels, stas, sids = [], [], [], []
for c in cats:
mris.extend(mris_d[c])
labels.extend(labels_d[c])
stas.extend(stas_d[c])
sids.extend(sids_d[c])
print('Dataset length: ', len(sids), len(mris), len(stas), len(labels))
# shuffle the dataset
index = np.random.permutation(np.arange(len(mris)))
mris, labels, stas, sids = np.asarray(mris), np.asarray(labels), np.asarray(stas), np.asarray(sids)
mris, labels, stas, sids = mris[index], labels[index], stas[index], sids[index]
return mris, labels, stas, sids