-
Notifications
You must be signed in to change notification settings - Fork 1
/
video_loader.py
233 lines (188 loc) · 7.52 KB
/
video_loader.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
from __future__ import print_function, absolute_import
import os
from PIL import Image
import numpy as np
import functools
from torchvision import transforms as T
import transforms as T
import sys
import torch
from torch import Tensor
from torch.utils.data import Dataset
import random
import torchvision.utils as vutil
# print(1)
# sys.setrecursionlimit(1000000000)
def read_image(img_path):
"""Keep reading image until succeed.
This can avoid IOError incurred by heavy IO process."""
got_img = False
while not got_img:
try:
img = Image.open(img_path).convert('RGB')
got_img = True
except IOError:
print("IOError incurred when reading '{}'. Will redo. Don't worry. Just chill.".format(img_path))
pass
return img
def get_default_video_loader():
image_loader = get_default_image_loader()
return functools.partial(video_loader, image_loader=image_loader)
def get_default_image_loader():
from torchvision import get_image_backend
if get_image_backend() == 'accimage':
return accimage_loader
else:
return pil_loader
def video_loader(img_paths, image_loader):
video = []
for image_path in img_paths:
if os.path.exists(image_path):
video.append(image_loader(image_path))
else:
return video
return video
def imge_loader(path):
from torchvision import get_image_backend
if get_image_backend() == 'accimage':
return accimage_loader(path)
else:
return pil_loader(path)
def accimage_loader(path):
try:
import accimage
return accimage.Image(path)
except IOError:
# Potentially a decoding problem, fall back to PIL.Image
return pil_loader(path)
def pil_loader(path):
with open(path,'rb') as f:
with Image.open(f) as img:
return img.convert('RGB')
def produce_out(imgs_path,seq_len, stride):
img_len = len(imgs_path)
frame_indices = list(range(img_len))
rand_end = max(0, img_len - seq_len * stride -1)
begin_index = random.randint(0, rand_end)
end_index = min(begin_index + seq_len * stride, img_len)
indices = frame_indices[begin_index:end_index]
re_indices= []
for i in range(0, seq_len * stride, stride):
add_arg = random.randint(0, stride-1)
re_indices.append(indices[i + add_arg])
re_indices = np.array(re_indices)
out = []
for index in re_indices:
out.append(imgs_path[int(index)])
return out
class VideoDataset(Dataset):
"""Video Person ReID Dataset.
Note batch data has shape (batch, seq_len, channel, height, width).
"""
sample_methods = ['evenly', 'random', 'all']
def __init__(self, dataset, seq_len=15, sample='evenly',
transform=None, max_seq_len=200, dataset_name="mars",
get_loader = get_default_video_loader,
):
self.dataset = dataset
self.seq_len = seq_len
self.sample = sample
self.transform = transform
self.max_seq_len = max_seq_len
self.dataset_name = dataset_name
self.loader = get_loader()
def __len__(self):
return len(self.dataset)
def __getitem__(self, index):
img_paths, pid, camid = self.dataset[index]
num = len(img_paths)
if self.sample == 'dense':
"""
Sample all frames in a video into a list of clips, each clip contains seq_len frames, batch_size needs to be set to 1.
This sampling strategy is used in test phase.
"""
frame_indices = list(range(num))
interval = num // self.seq_len
indices_list=[]
if num > self.seq_len:
for index in range(interval):
indices_list.append(frame_indices[index : index+interval * self.seq_len : interval])
else:
last_seq = frame_indices[0:]
for index in last_seq:
if len(last_seq) >= self.seq_len:
break
last_seq.append(index)
indices_list.append(last_seq)
imgs_list=[]
for indices in indices_list:
imgs = []
for index in indices:
index=int(index)
img_path = img_paths[index]
img = read_image(img_path)
if self.transform is not None:
img = self.transform(img)
img = img.unsqueeze(0)
imgs.append(img)
imgs = torch.cat(imgs, dim=0)
imgs_list.append(imgs)
if len(imgs_list) > self.max_seq_len:
sp = int(random.random() * (len(imgs_list) - self.max_seq_len))
ep = sp + self.max_seq_len
imgs_list = imgs_list[sp:ep]
imgs_array = torch.stack(imgs_list)
return imgs_array, pid, camid, img_paths[0]
elif self.sample == 'Begin_interval':
img_paths = list(img_paths)
interval = self.seq_len
num = self.seq_len - 1
if len(img_paths) >= interval * num + 1:
end_index = interval * num + 1
out = img_paths[0:end_index:interval]
elif len(img_paths) >= int(interval/2) * num + 1:
end_index = int(interval/2) * num + 1
out = img_paths[0:end_index:int(interval/2)]
elif len(img_paths) >= int(interval/4) * num + 1:
end_index = int(interval/4) * num + 1
out = img_paths[0:end_index:int(interval/4)]
elif len(img_paths) >= int(interval/8) * num + 1:
end_index = int(interval/8) * num + 1
out = img_paths[0:end_index:int(interval/8)]
else:
out = img_paths[0:interval]
while len(out) < interval:
for index in out:
if len(out) >= interval:
break
out.append(index)
clip = self.loader(out)
if self.transform is not None:
clip = [self.transform(img) for img in clip]
clip = torch.stack(clip, 0)
return clip, pid, camid, out
elif self.sample == 'Random_interval':
img_paths = list(img_paths)
stride = 8
if len(img_paths) >= self.seq_len * stride :
new_stride = stride
out = produce_out(img_paths, self.seq_len, new_stride)
elif len(img_paths) >= self.seq_len * int(stride/2):
new_stride = int(stride/2)
out = produce_out(img_paths, self.seq_len, new_stride)
elif len(img_paths) >= self.seq_len * int(stride/4):
new_stride = int(stride/4)
out = produce_out(img_paths, self.seq_len, new_stride)
elif len(img_paths) >= self.seq_len * int(stride/8):
new_stride = int(stride/8)
out = produce_out(img_paths, self.seq_len, new_stride)
else:
index = np.random.choice(len(img_paths), size=self.seq_len,replace=True)
index.sort()
out = [img_paths[index[i]] for i in range(self.seq_len)]
clip = self.loader(out)
clip = self.transform(clip)
clip = torch.stack(clip, 0)
return clip, pid, camid, out
else:
raise KeyError("Unknown sample method: {}. Expected one of {}".format(self.sample, self.sample_methods))