-
Notifications
You must be signed in to change notification settings - Fork 135
/
crop_vox.py
299 lines (242 loc) · 13.6 KB
/
crop_vox.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
import numpy as np
import pandas as pd
import imageio
import os
import subprocess
import warnings
import glob
import time
from util import bb_intersection_over_union, join, scheduler, crop_bbox_from_frames, save
from argparse import ArgumentParser
from skimage.transform import resize
warnings.filterwarnings("ignore")
DEVNULL = open(os.devnull, 'wb')
REF_FRAME_SIZE = 360
REF_FPS = 25
def extract_bbox(frame, refbbox, fa):
bboxes = fa.face_detector.detect_from_image(frame[..., ::-1])
if len(bboxes) != 0:
bbox = max([(bb_intersection_over_union(bbox, refbbox), tuple(bbox)) for bbox in bboxes])[1]
else:
bbox = np.array([0, 0, 0, 0, 0])
return np.maximum(np.array(bbox), 0)
def save_bbox_list(video_path, bbox_list):
f = open(os.path.join(args.bbox_folder, os.path.basename(video_path)[:-4] + '.txt'), 'w')
print("LEFT,TOP,RIGHT,BOT", file=f)
for bbox in bbox_list:
print("%s,%s,%s,%s" % tuple(bbox[:4]), file=f)
f.close()
def estimate_bbox(person_id, video_id, video_path, fa, args):
utterance = video_path.split('#')[1]
utterance = os.path.join(args.annotations_folder, person_id, video_id, utterance)
reader = imageio.get_reader(video_path)
bbox_list = []
d = pd.read_csv(utterance, sep='\t', skiprows=6)
frames = d['FRAME ']
try:
for i, frame in enumerate(reader):
if i >= len(frames):
break
val = d.iloc[i]
mult = frame.shape[0] / REF_FRAME_SIZE
frame = resize(frame, (REF_FRAME_SIZE, int(frame.shape[1] / mult)), preserve_range=True)
if args.dataset_version == 1:
x, y, w, h = val['X '], val['Y '], val['W '], val['H ']
else:
x, y, w, h = val['X '] * frame.shape[1], val['Y '] * frame.shape[0], val['W '] * frame.shape[1], val['H '] * frame.shape[0]
bbox = extract_bbox(frame, (x, y, x + w, y + h), fa)
bbox_list.append(bbox * mult)
except IndexError:
None
save_bbox_list(video_path, bbox_list)
def store(frame_list, tube_bbox, video_id, utterance, person_id, start, end, video_count, chunk_start, args):
out, final_bbox = crop_bbox_from_frames(frame_list, tube_bbox, min_frames=args.min_frames,
image_shape=args.image_shape, min_size=args.min_size,
increase_area=args.increase)
if out is None:
return []
start += round(chunk_start * REF_FPS)
end += round(chunk_start * REF_FPS)
name = (person_id + "#" + video_id + "#" + utterance + '#' + str(video_count).zfill(3) + ".mp4")
partition = 'test' if person_id in TEST_PERSONS else 'train'
save(os.path.join(args.out_folder, partition, name), out, args.format)
return [{'bbox': '-'.join(map(str, final_bbox)), 'start': start, 'end': end, 'fps': REF_FPS,
'video_id': '#'.join([video_id, person_id]), 'height': frame_list[0].shape[0],
'width': frame_list[0].shape[1], 'partition': partition}]
def crop_video(person_id, video_id, video_path, args):
utterance = video_path.split('#')[1]
bbox_path = os.path.join(args.bbox_folder, os.path.basename(video_path)[:-4] + '.txt')
reader = imageio.get_reader(video_path)
chunk_start = float(video_path.split('#')[2].split('-')[0])
d = pd.read_csv(bbox_path)
video_count = 0
initial_bbox = None
start = 0
tube_bbox = None
frame_list = []
chunks_data = []
try:
for i, frame in enumerate(reader):
bbox = np.array(d.iloc[i])
if initial_bbox is None:
initial_bbox = bbox
start = i
tube_bbox = bbox
if bb_intersection_over_union(initial_bbox, bbox) < args.iou_with_initial or len(
frame_list) >= args.max_frames:
chunks_data += store(frame_list, tube_bbox, video_id, utterance, person_id, start, i, video_count, chunk_start,
args)
video_count += 1
initial_bbox = bbox
start = i
tube_bbox = bbox
frame_list = []
tube_bbox = join(tube_bbox, bbox)
frame_list.append(frame)
except IndexError as e:
None
chunks_data += store(frame_list, tube_bbox, video_id, utterance, person_id, start, i + 1, video_count, chunk_start,
args)
return chunks_data
def download(video_id, args):
video_path = os.path.join(args.video_folder, video_id + ".mp4")
subprocess.call([args.youtube, '-f', "''best/mp4''", '--write-auto-sub', '--write-sub',
'--sub-lang', 'en', '--skip-unavailable-fragments',
"https://www.youtube.com/watch?v=" + video_id, "--output",
video_path], stdout=DEVNULL, stderr=DEVNULL)
return video_path
def split_in_utterance(person_id, video_id, args):
video_path = os.path.join(args.video_folder, video_id + ".mp4")
if not os.path.exists(video_path):
print("No video file %s found, probably broken link" % video_id)
return []
utterance_folder = os.path.join(args.annotations_folder, person_id, video_id)
utterance_files = sorted(os.listdir(utterance_folder))
utterances = [pd.read_csv(os.path.join(utterance_folder, f), sep='\t', skiprows=6) for f in
utterance_files]
chunk_names = []
for i, utterance in enumerate(utterances):
first_frame, last_frame = utterance['FRAME '].iloc[0], utterance['FRAME '].iloc[-1]
first_frame = round(first_frame / float(REF_FPS), 3)
last_frame = round(last_frame / float(REF_FPS), 3)
chunk_name = os.path.join(args.chunk_folder,
video_id + '#' + utterance_files[i] + '#' + str(first_frame) + '-' + str(
last_frame) + '.mp4')
chunk_names.append(chunk_name)
subprocess.call(['ffmpeg', '-y', '-i', video_path, '-qscale:v',
'5', '-r', '25', '-threads', '1', '-ss', str(first_frame), '-to', str(last_frame),
'-strict', '-2', '-deinterlace', chunk_name],
stdout=DEVNULL, stderr=DEVNULL)
return chunk_names
def run(params):
person_id, device_id, args = params
os.environ['CUDA_VISIBLE_DEVICES'] = device_id
# update the config options with the config file
if args.estimate_bbox:
import face_alignment
fa = face_alignment.FaceAlignment(face_alignment.LandmarksType._2D, flip_input=False)
video_folder = os.path.join(args.annotations_folder, person_id)
chunks_data = []
for video_id in os.listdir(video_folder):
intermediate_files = []
try:
if args.download:
video_path = download(video_id, args)
intermediate_files.append(video_path)
if args.split_in_utterance:
chunk_names = split_in_utterance(person_id, video_id, args)
intermediate_files += chunk_names
if args.estimate_bbox:
path = os.path.join(args.chunk_folder, video_id + '*.mp4')
for chunk in glob.glob(path):
while True:
try:
estimate_bbox(person_id, video_id, chunk, fa, args)
break
except RuntimeError as e:
if str(e).startswith('CUDA'):
print("Warning: out of memory, sleep for 1s")
time.sleep(1)
else:
print(e)
break
if args.crop:
path = os.path.join(args.chunk_folder, video_id + '*.mp4')
for chunk in glob.glob(path):
if not os.path.exists(os.path.join(args.bbox_folder, os.path.basename(chunk)[:-4] + '.txt')):
print ("BBox not found %s" % chunk)
continue
chunks_data += crop_video(person_id, video_id, chunk, args)
if args.remove_intermediate_results:
for file in intermediate_files:
if os.path.exists(file):
os.remove(file)
except Exception as e:
print (e)
return chunks_data
if __name__ == "__main__":
parser = ArgumentParser()
parser.add_argument("--dataset_version", default=1, type=int, choices=[1, 2], help='Version of Vox celeb dataset 1 or 2')
parser.add_argument("--iou_with_initial", type=float, default=0.25, help="The minimal allowed iou with inital bbox")
parser.add_argument("--image_shape", default=(256, 256), type=lambda x: tuple(map(int, x.split(','))),
help="Image shape")
parser.add_argument("--increase", default=0.1, type=float, help='Increase bbox by this amount')
parser.add_argument("--min_frames", default=64, type=int, help='Mimimal number of frames')
parser.add_argument("--max_frames", default=1024, type=int, help='Maximal number of frames')
parser.add_argument("--min_size", default=256, type=int, help='Minimal allowed size')
parser.add_argument("--format", default='.png', help='Store format (.png, .mp4)')
parser.add_argument("--annotations_folder", default='txt', help='Path to utterance annotations')
parser.add_argument("--video_folder", default='videos', help='Path to intermediate videos')
parser.add_argument("--chunk_folder", default='chunks', help="Path to folder with video chunks")
parser.add_argument("--bbox_folder", default='bbox', help="Path to folder with bboxes")
parser.add_argument("--out_folder", default='vox-png', help='Folder for processed dataset')
parser.add_argument("--chunks_metadata", default='vox-metadata.csv', help='File with metadata')
parser.add_argument("--youtube", default='./youtube-dl', help='Command for launching youtube-dl')
parser.add_argument("--workers", default=1, type=int, help='Number of parallel workers')
parser.add_argument("--device_ids", default="0", help="Names of the devices comma separated.")
parser.add_argument("--data_range", default=(0, 10000), type=lambda x: tuple(map(int, x.split('-'))), help="Range of ids for processing")
parser.add_argument("--no-download", dest="download", action="store_false", help="Do not download videos")
parser.add_argument("--no-split-in-utterance", dest="split_in_utterance", action="store_false",
help="Do not split videos in chunks")
parser.add_argument("--no-estimate-bbox", dest="estimate_bbox", action="store_false",
help="Do not estimate the bboxes")
parser.add_argument("--no-crop", dest="crop", action="store_false", help="Do not crop the videos")
parser.add_argument("--remove-intermediate-results", dest="remove_intermediate_results", action="store_true",
help="Remove intermediate videos")
parser.set_defaults(download=True)
parser.set_defaults(split_in_utterance=True)
parser.set_defaults(crop=True)
parser.set_defaults(estimate_bbox=True)
parser.set_defaults(remove_intermediate_results=False)
args = parser.parse_args()
if args.dataset_version == 1:
TEST_PERSONS = ['id' + str(i) for i in range(10270, 10310)]
else:
TEST_PERSONS = ['id07874', 'id00017', 'id00081', 'id09017', 'id08374', 'id04276', 'id03862', 'id00817', 'id00154',
'id02317', 'id06484', 'id07312', 'id03041', 'id05124', 'id03980', 'id05459', 'id04627', 'id08548',
'id01333', 'id02725', 'id05999', 'id06310', 'id08149', 'id04094', 'id08392', 'id02577', 'id01460',
'id02057', 'id08701', 'id00812', 'id00926', 'id03839', 'id06104', 'id07426', 'id08552', 'id01567',
'id03382', 'id02286', 'id03347', 'id08456', 'id02745', 'id00061', 'id01066', 'id03969', 'id06913',
'id01228', 'id02086', 'id08911', 'id01298', 'id06811', 'id07961', 'id04536', 'id01509', 'id01892',
'id08696', 'id06692', 'id01593', 'id01000', 'id01618', 'id04253', 'id04657', 'id04656', 'id03030',
'id01437', 'id02548', 'id01106', 'id04570', 'id05176', 'id05816', 'id00562', 'id02181', 'id07802',
'id03978', 'id04030', 'id03789', 'id04295', 'id00866', 'id07868', 'id04119', 'id01989', 'id07414',
'id01041', 'id03178', 'id04232', 'id03127', 'id06209', 'id03677', 'id04006', 'id05850', 'id02576',
'id05594', 'id01541', 'id05055', 'id07354', 'id01224', 'id03524', 'id02445', 'id07663', 'id05015',
'id07494', 'id04950', 'id04478', 'id02685', 'id02542', 'id05714', 'id02465', 'id05654', 'id05202',
'id00419', 'id03981', 'id04366', 'id07396', 'id02019', 'id01822', 'id06816', 'id07621', 'id07620', 'id04862']
if not os.path.exists(args.video_folder):
os.makedirs(args.video_folder)
if not os.path.exists(args.chunk_folder):
os.makedirs(args.chunk_folder)
if not os.path.exists(args.bbox_folder):
os.makedirs(args.bbox_folder)
if not os.path.exists(args.out_folder):
os.makedirs(args.out_folder)
for partition in ['test', 'train']:
if not os.path.exists(os.path.join(args.out_folder, partition)):
os.makedirs(os.path.join(args.out_folder, partition))
ids = set(os.listdir(args.annotations_folder))
ids_range = {'id' + str(num).zfill(5) for num in range(args.data_range[0], args.data_range[1])}
ids = sorted(list(ids.intersection(ids_range)))
scheduler(ids, run, args)