-
Notifications
You must be signed in to change notification settings - Fork 33
/
Copy pathmrflow_dataset.py
180 lines (131 loc) · 6.76 KB
/
mrflow_dataset.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
#! /usr/bin/env python2
import sys,os
import argparse
import subprocess
MRFLOW_HOME = os.environ['MRFLOW_HOME']
sys.path.append(MRFLOW_HOME)
import dataset_parameters
def generate_paths(dataset, arg):
""" Generate paths for either Sintel or Kitti
Requires the following environment variables to be set:
- SINTEL_HOME
- KITTI_HOME
- MRFLOW_SINTEL_INIT
- MRFLOW_KITTI_INIT
"""
if dataset == 'sintel':
testtrain, pas, seq, frame = arg.split(',')
frame = int(frame)
print('Calling Sintel preparation with')
print('\t TESTTRAIN = {}'.format(testtrain))
print('\t PASS = {}'.format(pas))
print('\t SEQ = {}'.format(seq))
print('\t FRAME = {}'.format(frame))
# Build paths for Sintel
SINTEL_HOME = os.environ['SINTEL_HOME']
PPPFLOW_SINTEL_INIT = os.environ['MRFLOW_SINTEL_INIT']
path_image_prev = os.path.join(SINTEL_HOME, testtrain, pas, seq, 'frame_{0:04d}.png'.format(frame-1))
path_image_current = os.path.join(SINTEL_HOME, testtrain, pas, seq, 'frame_{0:04d}.png'.format(frame))
path_image_next = os.path.join(SINTEL_HOME, testtrain, pas, seq, 'frame_{0:04d}.png'.format(frame+1))
# Flow from reference frame to adjacent frames
path_flow_fwd = os.path.join(PPPFLOW_SINTEL_INIT, 'flow', testtrain, pas, seq, 'frame_{0:04d}_fwd.flo'.format(frame))
path_flow_bwd = os.path.join(PPPFLOW_SINTEL_INIT, 'flow', testtrain, pas, seq, 'frame_{0:04d}_bwd.flo'.format(frame))
# Flow from adjacent frames back to reference frame
path_backflow_fwd = os.path.join(PPPFLOW_SINTEL_INIT, 'flow', testtrain, pas, seq, 'frame_{0:04d}_bwd.flo'.format(frame+1))
path_backflow_bwd = os.path.join(PPPFLOW_SINTEL_INIT, 'flow', testtrain, pas, seq, 'frame_{0:04d}_fwd.flo'.format(frame-1))
# Estimated rigidity
path_rigidity = os.path.join(PPPFLOW_SINTEL_INIT, 'rigidity', testtrain, pas, seq, 'frame_{0:04d}.png'.format(frame))
# Add GT regions if we are in training pass
if testtrain == 'training':
path_flow_fwd_gt = os.path.join(SINTEL_HOME, testtrain, 'flow', seq, 'frame_{0:04d}.flo'.format(frame))
path_rigidity_gt = os.path.join(SINTEL_HOME, testtrain, 'rigidity', seq, 'frame_{0:04d}.png'.format(frame))
else:
path_flow_fwd_gt = ''
path_rigidity_gt = ''
elif dataset == 'kitti':
testtrain, frame = arg.split(',')
frame = int(frame)
# KITTI
print('Calling KITTI preparation with')
print('\t TESTTRAIN = {}'.format(testtrain))
print('\t FRAME = {}'.format(frame))
# Build paths for Sintel
KITTI_HOME = os.environ['KITTI_HOME']
PPPFLOW_KITTI_INIT = os.environ['MRFLOW_KITTI_INIT']
# Hack for file layout
if testtrain == 'training':
testtrain_ = 'training'
else:
testtrain_ = 'testing'
path_image_prev = os.path.join(KITTI_HOME, testtrain_, 'image_2', '{0:06d}_09.png'.format(frame))
path_image_current = os.path.join(KITTI_HOME, testtrain_, 'image_2', '{0:06d}_10.png'.format(frame))
path_image_next = os.path.join(KITTI_HOME, testtrain_, 'image_2', '{0:06d}_11.png'.format(frame))
# Flow from reference frame to adjacent frames
path_flow_fwd = os.path.join(PPPFLOW_KITTI_INIT, 'flow', testtrain, '{0:06d}_10_fwd.flo'.format(frame))
path_flow_bwd = os.path.join(PPPFLOW_KITTI_INIT, 'flow', testtrain, '{0:06d}_10_bwd.flo'.format(frame))
# Flow from adjacent frames back to reference frame
path_backflow_bwd = os.path.join(PPPFLOW_KITTI_INIT, 'flow', testtrain, '{0:06d}_09_fwd.flo'.format(frame))
path_backflow_fwd = os.path.join(PPPFLOW_KITTI_INIT, 'flow', testtrain, '{0:06d}_11_bwd.flo'.format(frame))
# Estimated rigidity
path_rigidity = os.path.join(PPPFLOW_KITTI_INIT, 'rigidity', testtrain, '{0:06d}_10.png'.format(frame))
# Add GT regions if we are in training pass
if testtrain == 'training':
path_flow_fwd_gt = os.path.join(KITTI_HOME, testtrain, 'flow_occ', '{0:06d}_10.png'.format(frame))
path_rigidity_gt = os.path.join(KITTI_HOME, testtrain, 'rigidity_generated', '{0:06d}_10.png'.format(frame))
else:
path_flow_fwd_gt = ''
path_rigidity_gt = ''
paths = {
'--flow_fwd': path_flow_fwd,
'--flow_bwd': path_flow_bwd,
'--backflow_fwd': path_backflow_fwd,
'--backflow_bwd': path_backflow_bwd,
'--rigidity': path_rigidity
}
if path_flow_fwd_gt:
paths['--flow_fwd_gt'] = path_flow_fwd_gt
if path_rigidity_gt:
paths['--rigidity_gt'] = path_rigidity_gt
paths_images = [path_image_prev, path_image_current, path_image_next]
return paths,paths_images
def main():
parser = argparse.ArgumentParser()
parser.add_argument('dataset', type=str, help='Dataset to use (sintel/kitti)')
parser.add_argument('token', type=str, help='Token determining the frame.\n For KITTI, please give as {training/test},frame.\n For Sintel, give as {training/test},pass,seq,frame.')
parser.add_argument('args', nargs=argparse.REMAINDER)
args = parser.parse_args()
paths,paths_images = generate_paths(args.dataset,args.token)
if args.dataset == 'kitti':
testtrain, frame = args.token.split(',')
params_default = dataset_parameters.kitti_parameters
params_default['tempdir'] = os.path.join('data_seqs', testtrain, '{0:06d}'.format(int(frame)))
elif args.dataset == 'sintel':
testtrain, pas, seq, frame = args.token.split(',')
params_default = dataset_parameters.sintel_parameters
params_default['tempdir'] = os.path.join('data_seqs', testtrain, pas, seq, 'frame_{0:04d}'.format(int(frame)))
# If tempdir does not exist yet, create it.
if not os.path.isdir(params_default['tempdir']):
os.makedirs(params_default['tempdir'])
# Set up params to call mr-flow with
args_mrflow = {}
for k,v in params_default.items():
args_mrflow['--' + k] = str(v)
for k,v in paths.items():
args_mrflow[k] = v
remainder_args = zip(args.args[::2],args.args[1::2])
for k,v in remainder_args:
args_mrflow[k] = v
args_mrflow_array = []
for k,v in args_mrflow.items():
args_mrflow_array.append(k)
args_mrflow_array.append(v)
args_mrflow_array.append(paths_images[0])
args_mrflow_array.append(paths_images[1])
args_mrflow_array.append(paths_images[2])
print('Calling MR-Flow with arguments: ')
for k,v in zip(args_mrflow_array[::2],args_mrflow_array[1::2]):
print('\t{}\t:\t{}'.format(k,v))
print('')
subprocess.call(['python', 'mrflow.py',] + args_mrflow_array)
if __name__ == '__main__':
main()