-
Notifications
You must be signed in to change notification settings - Fork 19
/
Copy pathmain.py
204 lines (180 loc) · 8.5 KB
/
main.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
# Software License Agreement (BSD License)
#
# Copyright (c) 2019, Zerong Zheng ([email protected])
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of the <organization> nor the
# names of its contributors may be used to endorse or promote products
# derived from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
# DISCLAIMED. IN NO EVENT SHALL <COPYRIGHT HOLDER> BE LIABLE FOR ANY
# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
# ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
from __future__ import print_function, absolute_import, division
import os
import numpy as np
import cv2 as cv
import argparse
import random
import string
import shutil
from subprocess import call
import pynvml
import time
pynvml.nvmlInit()
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument('--img_file', type=str, required=True, help='path to image file')
parser.add_argument('--out_dir', type=str, required=True, help='output directory')
return parser.parse_args()
def waitgpu(empty_thres_duration=7):
empty_flag = 0
while True:
handle = pynvml.nvmlDeviceGetHandleByIndex(0)
meminfo = pynvml.nvmlDeviceGetMemoryInfo(handle)
usage_percent = float(meminfo.used)/float(meminfo.total)
if usage_percent < 0.1:
if empty_flag >= empty_thres_duration: # empty for 5 second
break
empty_flag += 1
time.sleep(1)
continue
empty_flag = 0
print('GPU is busy right now....waiting....')
print('meminfo.used/meminfo.total = %f' % usage_percent)
time.sleep(np.random.randint(5, 15))
def detect_human(fname, out_dir):
""" obtains bounding box of the subject in the input image"""
waitgpu()
print('\n\nStep 1. Human Detection RCNN')
# generate a temporal script to call RCNN
shutil.copy('./detect_human.py', './AlphaPose/human-detection/tools/')
temp_shname = ''.join(random.sample(string.ascii_letters + string.digits, 8)) + '.sh'
temp_shname = os.path.join('./', temp_shname)
with open(temp_shname, 'w') as fp:
fp.write('#!/usr/local/bin/bash\n')
fp.write('cd ./AlphaPose/human-detection/tools\n')
fp.write('python2 detect_human.py --img_file %s --out_dir %s\n'
% (fname, out_dir))
fp.write('cd ../../../\n')
call(['sh', temp_shname])
os.remove(temp_shname)
# os.remove('./AlphaPose/human-detection/tools/detect_human.py')
def crop_or_pad_img(fname, out_dir):
""" crops or pads the original image to make the subject located at the center
of the image and occupy 90% of the image
"""
print('\n\nStep 2. Image cropping or padding')
img_dir, img_name = os.path.split(img_fname)
with open(os.path.join(out_dir, img_name + '.bbox.txt'), 'r') as fp:
lines = fp.readlines()
if lines[-1] == '':
lines = lines[:-1]
if len(lines) > 1:
print('Warning: More than 1 bounding boxes are detected. '
'Only the first one is used.')
entries = lines[0].split(' ')
xmin, ymin = int(entries[0]), int(entries[1])
xmax, ymax = int(entries[2]), int(entries[3])
x_center = int((xmin+xmax)/2)
y_center = int((ymin+ymax)/2)
edge_len = int(max(xmax-xmin, ymax-ymin) * 1.2)
edge_len_half = int(edge_len/2)
img = cv.imread(fname)
cv.imwrite(os.path.join(out_dir, img_name[:-4]+'_orig.png'), img)
if len(img.shape) == 2:
img = cv.cvtColor(img, cv.COLOR_GRAY2RGB)
h, w = img.shape[0], img.shape[1]
img_pad = np.zeros((3*h, 3*w, 3), dtype=np.uint8)
img_pad[h:(h*2), w:(w*2), :] = img
crop_tl = (h+y_center-edge_len_half, w+x_center-edge_len_half)
crop_dr = (h+y_center+edge_len_half, w+x_center+edge_len_half)
img_crop = img_pad[crop_tl[0]:crop_dr[0], crop_tl[1]:crop_dr[1], :]
cv.imwrite(os.path.join(out_dir, img_name), img_crop)
cv.imwrite(os.path.join(out_dir, img_name), img_crop)
def infer_smpl_and_pose(fname, out_dir):
waitgpu()
print('\n\nStep 3a Body model estimation using HMR. ')
shutil.copy('./infer_smpl.py', './hmr/')
temp_shname = ''.join(random.sample(string.ascii_letters + string.digits, 8)) + '.sh'
temp_shname = os.path.join('./', temp_shname)
with open(temp_shname, 'w') as fp:
fp.write('#!/usr/local/bin/bash\n')
fp.write('cd ./hmr/\n')
fp.write('python2 infer_smpl.py --img_path %s --out_dir %s\n' % (fname, out_dir))
fp.write('cd ../\n')
call(['sh', temp_shname])
os.remove(temp_shname)
# os.remove('./hmr/infer_smpl.py')
print('\n\nStep 3b Pose estimation using AlphaPose')
img_dir, img_name = os.path.split(img_fname)
tmp_folder = ''.join(random.sample(string.ascii_letters + string.digits, 8))
os.mkdir(os.path.join('./AlphaPose/examples', tmp_folder))
os.mkdir(os.path.join('./AlphaPose/examples', tmp_folder, 'demo'))
os.mkdir(os.path.join('./AlphaPose/examples', tmp_folder, 'results'))
call(['cp', os.path.join(out_dir, img_name),
os.path.join('./AlphaPose/examples', tmp_folder, 'demo/1.jpg')])
call(['./AlphaPose/run.sh', '--indir', os.path.join('./examples', tmp_folder, 'demo'),
'--outdir', os.path.join('./examples', tmp_folder, 'results'), '--vis'])
call(['mv', os.path.join('./AlphaPose/examples', tmp_folder, 'results/POSE/pred.txt'),
os.path.join(out_dir, img_name+'.joints.txt')])
call(['mv', os.path.join('./AlphaPose/examples', tmp_folder, 'results/POSE/scores.txt'),
os.path.join(out_dir, img_name+'.joint_scores.txt')])
call(['rm', '-r', os.path.join('./AlphaPose/examples', tmp_folder)])
print('\n\nStep 3c Image segmentation')
shutil.copy('./segment_by_parsing.py', './LIP_JPPNet/')
temp_shname = ''.join(random.sample(string.ascii_letters + string.digits, 8)) + '.sh'
temp_shname = os.path.join('./', temp_shname)
with open(temp_shname, 'w') as fp:
fp.write('#!/usr/local/bin/bash\n')
fp.write('cd ./LIP_JPPNet/\n')
fp.write('python2 segment_by_parsing.py --img_file %s --out_dir %s\n' % (fname, out_dir))
fp.write('cd ../\n')
call(['sh', temp_shname])
os.remove(temp_shname)
# os.remove('./LIP_JPPNet/segment_by_parsing.py')
def optimize_smpl(fname, out_dir):
print('\n\nStep 4 SMPL model optimization')
shutil.copy('./fit_3d_accurate.py', './smplify_public/code/')
temp_shname = ''.join(random.sample(string.ascii_letters + string.digits, 8)) + '.sh'
temp_shname = os.path.join('./', temp_shname)
with open(temp_shname, 'w') as fp:
fp.write('#!/usr/local/bin/bash\n')
fp.write('cd ./smplify_public/code\n')
fp.write('python2 fit_3d_accurate.py --img_file %s --out_dir %s\n' % (fname, out_dir))
fp.write('cd ../../\n')
call(['sh', temp_shname])
os.remove(temp_shname)
# os.remove('smplify_public/code/fit_3d_accurate.py')
def main(img_fname, out_dir):
print('image file: ' + img_fname)
print('output directory: ' + out_dir)
if not os.path.isfile(img_fname):
raise IOError('Image file does not exist!')
if not os.path.exists(out_dir):
os.mkdir(out_dir)
detect_human(img_fname, out_dir)
crop_or_pad_img(img_fname, out_dir)
infer_smpl_and_pose(img_fname, out_dir)
optimize_smpl(img_fname, out_dir)
if __name__ == '__main__':
args = parse_args()
img_fname = args.img_file
out_dir = args.out_dir
img_fname = os.path.abspath(img_fname)
out_dir = os.path.abspath(out_dir)
main(img_fname, out_dir)