-
Notifications
You must be signed in to change notification settings - Fork 0
/
data_generation.py
329 lines (251 loc) · 9.95 KB
/
data_generation.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
#-*-coding: utf-8-*-
from text_generation import *
from config import Config
import os
import json
import csv
import shutil
import random
from xml.etree.ElementTree import parse
import numpy as np
import pandas as pd
from collections import OrderedDict
def extractImdir(data_dir):
img_dir_aligned_full = []
img_dir_flip_full = []
img_dir_aligned = []
img_dir_flip = []
dir_image = os.path.join(data_dir, 'cohn-kanade-images')
files = sorted(os.listdir(dir_image))
for name in files:
dir_image_1 = os.path.join(dir_image, name)
image_files = sorted(os.listdir(dir_image_1))
for child_name in image_files:
dir_image_2 = os.path.join(dir_image_1, child_name)
if os.path.isdir(dir_image_2):
image_files_2 = sorted(os.listdir(dir_image_2))
dir_image_aligned = os.path.join(dir_image_2, image_files_2[-2])
dir_image_flip = os.path.join(dir_image_2, image_files_2[-1])
img_dir_aligned_full.append(dir_image_aligned)
img_dir_flip_full.append(dir_image_flip)
# _, emo_dir, _ = extractAUdir(data_dir)
# for i in range(len(img_dir_aligned_full)):
# if any(img_dir_aligned_full[i][-29:-12] in img_string for img_string in emo_dir):
# img_dir_aligned.append(img_dir_aligned_full[i])
# img_dir_flip.append(img_dir_flip_full[i])
return img_dir_aligned_full, img_dir_flip_full
def extractImdir_rough(data_dir):
dir_full = []
dir_image = os.path.join(data_dir, 'cohn-kanade-images')
files = sorted(os.listdir(dir_image))
for name in files:
dir_image_1 = os.path.join(dir_image, name)
image_files = sorted(os.listdir(dir_image_1))
for child_name in image_files:
dir_image_2 = os.path.join(dir_image_1, child_name)
if os.path.isdir(dir_image_2):
image_files_2 = sorted(os.listdir(dir_image_2))
for img in image_files_2:
dir_image_full = os.path.join(dir_image_2, img)
if dir_image_full[-3:] == 'png' and not dir_image_full[-11:-4] == 'aligned' and not dir_image_full[
-11:-4] == 'flipped':
dir_full.append(dir_image_full)
return dir_full
def extractAUdir(data_dir):
'''extract Action Unit directory (list)'''
Genderdir = []
Emotiondir = []
AUdir = []
### mmi dataset has no emotion ###
if 'mmi' in data_dir:
files = sorted(os.listdir(data_dir))
files.sort(key=int) # sort
for name in files:
fn = os.path.join(data_dir, name)
aucs = filter(lambda x: 'aucs' in x, os.listdir(fn))[0]
fn_aucs = os.path.join(fn, aucs)
AUdir.append(fn_aucs)
elif 'CK+' in data_dir:
emo_fn = os.path.join(data_dir, 'Emotion')
au_fn = os.path.join(data_dir, 'FACS')
files = sorted(os.listdir(emo_fn))
for name in files:
emo_fn_full = os.path.join(emo_fn, name)
au_fn_full = os.path.join(au_fn, name)
gd_fn_full = os.path.join(au_fn_full, 'gender.txt')
sess = sorted(os.listdir(au_fn_full))
for aucs_dir in sess:
if not 'gender' in aucs_dir:
au_fn_aucs_dir = os.path.join(au_fn_full, aucs_dir)
emo_fn_aucs_dir = os.path.join(emo_fn_full, aucs_dir)
# if os.listdir(emo_fn_aucs_dir):
Genderdir.append(gd_fn_full)
aucs = os.listdir(au_fn_aucs_dir)[0]
fn_aucs = os.path.join(au_fn_aucs_dir, aucs)
AUdir.append(fn_aucs)
if not os.path.exists(emo_fn_aucs_dir):
fn_emos = '.'
elif not os.listdir(emo_fn_aucs_dir):
fn_emos = '.'
else:
emos = os.listdir(emo_fn_aucs_dir)[0]
fn_emos = os.path.join(emo_fn_aucs_dir, emos)
Emotiondir.append(fn_emos)
return Genderdir, Emotiondir, AUdir
def extractAU(data_dir):
''' extract Action Unit data with intensity '''
type_dir = data_dir[-3:]
if type_dir == 'xml':
tree = parse(data_dir)
root = tree.getroot()
AU = pd.DataFrame()
for ActionUnit in root.findall('ActionUnit'):
numAU = ActionUnit.attrib
for marker in ActionUnit:
dict = marker.attrib
dict.update(numAU)
dict['Intensity'] = dict.pop('Type')
dict['ActionUnit'] = dict.pop('Number')
AU = AU.append(dict, ignore_index=True)
elif type_dir == 'txt':
AU = pd.read_csv(data_dir, delim_whitespace=True, names=["ActionUnit", "Intensity"])
else:
return None
return AU
def extractInfo(data_dir, name='Gender'):
'''gender or emotion information extraction'''
type_dir = data_dir[-3:]
if type_dir == 'txt':
Emo = pd.read_csv(data_dir, delim_whitespace=True, names=[name])
return Emo
def shuffleDataset(dir):
img_aligned, img_flip = extractImdir(dir)
gen, emo, au = extractAUdir(dir)
shuffleAU = list(zip(img_aligned, img_flip, gen, emo, au))
random.shuffle(shuffleAU)
img_aligned, img_flip, gen, emo, au = zip(*shuffleAU)
return [img_aligned, img_flip, gen, emo, au]
def creatDataset(data, num, config):
img_aligned, img_flip, gen, emo, au = data
# img_aligned, img_flip = extractImdir(dir)
# gen, emo, au = extractAUdir(dir)
anns, imgs, genders, emos, aus = [], [], [], [], []
ann = OrderedDict()
for i in range(len(emo)-50): # last 100 for validation set
df = summary([gen[i], emo[i], au[i]])
text = generateText(df, numtext=num*2, orderS=config.orderS, orderG=config.orderG)
for j in range(num):
anns.append(text[j])
imgs.append(img_aligned[i])
genders.append(gen[i])
emos.append(emo[i])
aus.append(au[i])
for k in range(num):
anns.append(text[k+num])
imgs.append(img_flip[i])
genders.append(gen[i])
emos.append(emo[i])
aus.append(au[i])
ann['caption'] = anns
ann['image_file'] = imgs
ann['gender'] = genders
ann['emotion'] = emos
ann['action_unit'] = aus
ann = pd.DataFrame.from_dict(ann)
data_dir = dir + 'annotation.csv'
if os.path.exists(data_dir):
os.remove(data_dir)
ann.to_csv('annotation.csv')
return ann
def creatDataset_test(data, num, config):
img_aligned, img_flip, gen, emo, au = data
# img_aligned, img_flip = extractImdir(dir)
# gen, emo, au = extractAUdir(dir)
anns, imgs, genders, emos, aus = [], [], [], [], []
ann = OrderedDict()
for q in range(50): # last 100 for validation set
i=len(emo)-50+q
df = summary([gen[i], emo[i], au[i]])
text = generateText(df, numtext=num*2, orderS=config.orderS, orderG=config.orderG)
for j in range(num):
anns.append(text[j])
imgs.append(img_aligned[i])
genders.append(gen[i])
emos.append(emo[i])
aus.append(au[i])
for k in range(num):
anns.append(text[k+num])
imgs.append(img_flip[i])
genders.append(gen[i])
emos.append(emo[i])
aus.append(au[i])
ann['caption'] = anns
ann['image_file'] = imgs
ann['gender'] = genders
ann['emotion'] = emos
ann['action_unit'] = aus
ann = pd.DataFrame.from_dict(ann)
data_dir = dir + 'annotation_test.csv'
if os.path.exists(data_dir):
os.remove(data_dir)
ann.to_csv('annotation_test.csv')
return ann
def summary(data_dirs):
'''
:param data_dirs: [gender directory, emotion directory, action unit directory]
:return:
'''
[gen, emo, au] = data_dirs
dfgen = extractInfo(gen, 'Gender')
if emo == '.':
data = {'Emotion':['0']}
dfemo = pd.DataFrame(data)
else:
dfemo = extractInfo(emo, 'Emotion')
dfau = extractAU(au)
df = pd.concat([dfgen, dfemo], axis=1)
df = pd.concat([df, dfau], axis=1)
df['Emotion'] = df['Emotion'].fillna(df['Emotion'][0])
df['Gender'] = df['Gender'].fillna(df['Gender'][0])
# df = df.fillna("") # for good-looking
return df
if __name__ == '__main__':
dir = './CK+'
# dir2 = './mmi-facial-expression-database/Sessions'
# dir3 = './mmi-facial-expression-database2/
img_aligned, img_flip = extractImdir(dir)
gen, emo, au = extractAUdir(dir)
# data = shuffleDataset(dir)
# print(data)
# a = pd.read_csv('directory.csv')
# print(a)
# print(list(a.iloc[0][0:2]))
# print(list(a.iloc[0][2:5]))
full = []
for i in range(len(emo)):
print('Image directory (aligned) : {}'.format(img_aligned[i]))
print('Image directory (flipped) : {}'.format(img_flip[i]))
print('Gender directory: {}'.format(gen[i]))
print('Emotion directory: {}'.format(emo[i]))
print('ActionUnit directory: {}'.format(au[i]))
data_dir = [gen[i], emo[i], au[i]]
df = summary(data_dir)
print(np.asarray(df['ActionUnit'])-1)
# #print(set(df['Gender']), set(df['Emotion']), df['ActionUnit'], df['Intensity'])
#
# full.append(df)
#
# full_df = pd.concat(full)
# print(full_df)
# print(set(full_df['Gender']))
# print(set(full_df['Emotion']))
# print(set(full_df['ActionUnit']))
# print(set(full_df['Intensity']))
################ Example ################
# Gender Emotion ActionUnit Intensity
# 0 0 4 1.0 4.0
# 1 2.0 4.0
# 2 4.0 2.0
# 3 5.0 4.0
# 4 20.0 3.0
# 5 25.0 1.0