-
Notifications
You must be signed in to change notification settings - Fork 2
Expand file tree
/
Copy pathdatasets.py
More file actions
213 lines (194 loc) · 11.5 KB
/
Copy pathdatasets.py
File metadata and controls
213 lines (194 loc) · 11.5 KB
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
import torch
import torchvision.transforms as transforms
import os
from os.path import join
import glob
import cv2
import numpy as np
import pandas as pd
from math import floor
from PIL import Image
import torch.utils.data as data
from RIFE.interpolate_img import interpolate
class ME_dataset(data.Dataset):
def __init__(self, transform=None, train=True, val_subject=1, dataset='smic', datadir='inputs', combined = False):
self.transform = transform
self.val_subject = val_subject
self.train = train
self.dataset = dataset
self.data_folder = join(datadir,dataset,'LOF')
self.labels = ('negative','positive','surprise')
if combined:
data_gt = pd.read_csv(join(datadir,'combined_3class_gt.csv'), header=None, names=['Dataset','Subject','Filename','Class'])
elif dataset == 'casme2':
self.labels = ('disgust', 'happiness', 'repression', 'surprise', 'others')
elif dataset == 'samm':
self.labels = ('Other', 'Anger', 'Contempt', 'Happiness', 'Surprise')
# elif dataset == 'ck+':
# self.labels = ('neutral', 'anger', 'contempt', 'disgust', 'fear', 'happy', 'sadness', 'surprise')
# self.ftype = '.png'
data_info = pd.read_csv(join(datadir,dataset,'coding.csv'))
# label indexing, {'negative': array(0}, ...}
self.label_index = {label : i for i, label in enumerate(self.labels)}
self.subject_num = data_info.Subject.nunique()
if combined:
self.subject_num = data_gt.Subject.nunique()
image_list = glob.glob(join(self.data_folder,'*','*','*'+'.jpg'), recursive=True)
if self.dataset == 'smic':
image_list = glob.glob(join(self.data_folder,'*','*','*','*'+'.bmp'), recursive=True)
self.train_image = image_list.copy()
self.val_image = []
self.val_label = {}
self.val_images = {}
self.train_images = {}
self.train_label = {}
val_imgs = {}
train_imgs = {}
if self.dataset == 'casme2':
for image in image_list:
sub = data_info[data_info['Subject'].isin([int(image.split('/')[-3][3:])])]
ep = sub[sub['Filename'].isin([image.split('/')[-2]])]
if combined:
sub = data_gt[data_gt['Subject'].isin([image.split('/')[-3]])]
ep = sub[sub['Filename'].isin([image.split('/')[-2]])]
if sub.empty or ep.empty:
self.train_image.remove(image)
elif str(ep.Subject.values[0]) in [str(self.val_subject), 'sub'+str(self.val_subject).zfill(2)]:
if combined:
self.val_label[image] = ep.Class.values[0]
if ep.Subject.values[0]+ep.Filename.values[0] not in val_imgs.keys():
val_imgs[ep.Subject.values[0]+ep.Filename.values[0]] = []
val_imgs[ep.Subject.values[0]+ep.Filename.values[0]].append(image)
elif ep.Emotion.values[0] in self.labels:
self.val_label[image] = self.label_index[ep.Emotion.values[0]]
if 'sub'+str(ep.Subject.values[0]).zfill(2)+ep.Filename.values[0] not in val_imgs.keys():
val_imgs['sub'+str(ep.Subject.values[0]).zfill(2)+ep.Filename.values[0]] = []
val_imgs['sub'+str(ep.Subject.values[0]).zfill(2)+ep.Filename.values[0]].append(image)
self.train_image.remove(image)
else:
if combined:
self.train_label[image] = ep.Class.values[0]
if ep.Subject.values[0]+ep.Filename.values[0] not in train_imgs.keys():
train_imgs[ep.Subject.values[0]+ep.Filename.values[0]] = []
train_imgs[ep.Subject.values[0]+ep.Filename.values[0]].append(image)
elif ep.Emotion.values[0] in self.labels:
self.train_label[image] = self.label_index[ep.Emotion.values[0]]
if 'sub'+str(ep.Subject.values[0]).zfill(2)+ep.Filename.values[0] not in train_imgs.keys():
train_imgs['sub'+str(ep.Subject.values[0]).zfill(2)+ep.Filename.values[0]] = []
train_imgs['sub'+str(ep.Subject.values[0]).zfill(2)+ep.Filename.values[0]].append(image)
else:
self.train_image.remove(image)
elif self.dataset == 'samm':
for image in image_list:
sub = data_info[data_info['Subject'].isin([int(image.split('/')[-3])])]
ep = sub[sub['Filename'].isin([image.split('/')[-2]])]
if combined:
sub = data_gt[data_gt['Subject'].isin([str(int(image.split('/')[-3]))])]
ep = sub[sub['Filename'].isin([image.split('/')[-2]])]
if sub.empty or ep.empty:
self.train_image.remove(image)
elif str(ep.Subject.values[0]) in [str(self.val_subject), str(self.val_subject).zfill(3)]:
if combined:
self.val_label[image] = ep.Class.values[0]
if str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0] not in val_imgs.keys():
val_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]] = []
val_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]].append(image)
elif ep.Emotion.values[0] in self.labels:
self.val_label[image] = self.label_index[ep.Emotion.values[0]]
if str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0] not in val_imgs.keys():
val_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]] = []
val_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]].append(image)
self.train_image.remove(image)
else:
if combined:
self.train_label[image] = ep.Class.values[0]
if str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0] not in train_imgs.keys():
train_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]] = []
train_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]].append(image)
elif ep.Emotion.values[0] in self.labels:
self.train_label[image] = self.label_index[ep.Emotion.values[0]]
if str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0] not in train_imgs.keys():
train_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]] = []
train_imgs[str(ep.Subject.values[0]).zfill(3)+ep.Filename.values[0]].append(image)
else:
self.train_image.remove(image)
elif self.dataset == 'smic':
for image in image_list:
sub = data_info[data_info['Subject'].isin([image.split('/')[-4]])]
ep = sub[sub['Filename'].isin([image.split('/')[-2]])]
if combined:
sub = data_gt[data_gt['Subject'].isin([image.split('/')[-4]])]
ep = sub[sub['Filename'].isin([image.split('/')[-2]])]
if sub.empty or ep.empty:
self.train_image.remove(image)
elif str(ep.Subject.values[0]) in [str(self.val_subject), 's'+str(self.val_subject)]:
if combined:
self.val_label[image] = ep.Class.values[0]
if ep.Subject.values[0]+ep.Filename.values[0] not in val_imgs.keys():
val_imgs[ep.Subject.values[0]+ep.Filename.values[0]] = []
val_imgs[ep.Subject.values[0]+ep.Filename.values[0]].append(image)
elif ep.Emotion.values[0] in self.labels:
self.val_label[image] = self.label_index[ep.Emotion.values[0]]
if str(ep.Subject.values[0])+ep.Filename.values[0] not in val_imgs.keys():
val_imgs[str(ep.Subject.values[0])+ep.Filename.values[0]] = []
val_imgs[str(ep.Subject.values[0])+ep.Filename.values[0]].append(image)
self.train_image.remove(image)
else:
if combined:
self.train_label[image] = ep.Class.values[0]
if ep.Subject.values[0]+ep.Filename.values[0] not in train_imgs.keys():
train_imgs[ep.Subject.values[0]+ep.Filename.values[0]] = []
train_imgs[ep.Subject.values[0]+ep.Filename.values[0]].append(image)
elif ep.Emotion.values[0] in self.labels:
self.train_label[image] = self.label_index[ep.Emotion.values[0]]
if str(ep.Subject.values[0])+ep.Filename.values[0] not in train_imgs.keys():
train_imgs[str(ep.Subject.values[0])+ep.Filename.values[0]] = []
train_imgs[str(ep.Subject.values[0])+ep.Filename.values[0]].append(image)
else:
self.train_image.remove(image)
# elif self.dataset == 'ck+':
# for image in image_list:
# sub = data_info[data_info['Subject'].isin([image.split('/')[-3]])]
# ep = sub[sub['Filename'].isin([image.split('/')[-2]])]
# self.train_label[image] = self.label_index[ep.Emotion.values[0]]
# if ep.Subject.values[0]+str(ep.Filename.values[0]) not in train_imgs.keys():
# train_imgs[ep.Subject.values[0]+str(ep.Filename.values[0])] = []
# train_imgs[ep.Subject.values[0]+str(ep.Filename.values[0])].append(image)
for i, key in enumerate(val_imgs.keys(), 0):
self.val_images[i] = val_imgs[key]
for i, key in enumerate(train_imgs.keys(), 0):
self.train_images[i] = train_imgs[key]
def __getitem__(self, index):
"""
Args:
index (int): Index
Returns:
tuple: (image, target, subject) where target is index of the target class, subject is the subject number
"""
img, target = [], []
if self.train:
for image in self.train_images[index]:
target.append(self.train_label[image])
# doing this so that it is consistent with all other datasets
# to return a PIL Image
im = Image.open(image)
if self.transform is not None:
im = self.transform(im)
img.append(im)
else:
for image in self.val_images[index]:
target.append(self.val_label[image])
# doing this so that it is consistent with all other datasets
# to return a PIL Image
im = Image.open(image)
if self.transform is not None:
im = self.transform(im)
img.append(im)
img = torch.stack(img)
target = torch.LongTensor(target)
return img, target
def __len__(self):
if self.train:
return len(self.train_images)
else:
return len(self.val_images)