forked from open-mmlab/mmsegmentation
-
Notifications
You must be signed in to change notification settings - Fork 0
/
video_slice.py
168 lines (135 loc) · 5.13 KB
/
video_slice.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
from mmseg.apis import inference_segmentor, init_segmentor
from mmpose.apis import (get_track_id, inference_top_down_pose_model,
init_pose_model, vis_pose_tracking_result)
from mmdet.apis import inference_detector, init_detector
import mmcv
import os
import cv2
from argparse import ArgumentParser
import pathlib
from datetime import datetime
import json
zoo = {
'pspnet': {
'config':
'configs/pspnet/pspnet_r101-d8_512x512_40k_voc12aug.py',
'checkpoint':
'checkpoints/pspnet_r101-d8_512x512_40k_voc12aug_20200613_161222-bc933b18.pth'
},
'deeplabv3': {
'config':
'configs/deeplabv3/deeplabv3_r101-d8_512x512_40k_voc12aug.py',
'checkpoint':
'checkpoints/deeplabv3_r101-d8_512x512_40k_voc12aug_20200613_161432-0017d784.pth'
},
'dlc59': {
'config':
'configs/deeplabv3/deeplabv3_r101-d8_480x480_80k_pascal_context_59.py',
'checkpoint':
'checkpoints/deeplabv3_r101-d8_480x480_80k_pascal_context_59_20210416_113002-26303993.pth'
},
'psanet': {
'config':
'configs/psanet/psanet_r101-d8_512x512_20k_voc12aug.py',
'checkpoint':
'checkpoints/psanet_r101-d8_512x512_20k_voc12aug_20200617_110624-946fef11.pth'
}
}
def process_mmdet_results(mmdet_results, cat_id=1):
"""Process mmdet results, and return a list of bboxes.
:param mmdet_results:
:param cat_id: category id (default: 1 for human)
:return: a list of detected bounding boxes
"""
if isinstance(mmdet_results, tuple):
det_results = mmdet_results[0]
else:
det_results = mmdet_results
bboxes = det_results[cat_id - 1]
person_results = []
for bbox in bboxes:
person = {}
person['bbox'] = bbox
person_results.append(person)
return person_results
def get_model_files(zoo_id):
return zoo[zoo_id]['config'], zoo[zoo_id]['checkpoint']
def init_video(args):
cap = cv2.VideoCapture(args.video_path)
fps = cap.get(cv2.CAP_PROP_FPS)
print('video fps = ', fps)
size = (int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)),
int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)))
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
videoWriter = cv2.VideoWriter(
os.path.join(args.out_video_root,
f'vis_{os.path.basename(args.video_path)}'), fourcc, fps,
size)
return cap, fps, videoWriter
# optional
return_heatmap = False
# e.g. use ('backbone', ) to return backbone feature
output_layer_names = None
def main():
parser = ArgumentParser()
parser.add_argument('--video-path', type=str, help='Video path')
parser.add_argument(
'--out-video-root', type=str, default='.', help='Output directory')
parser.add_argument(
'--show',
action='store_true',
default=False,
help='whether to show visualizations.')
args = parser.parse_args()
save_out_video = True
config_file, checkpoint_file = get_model_files('psanet')
# build the model from a config file and a checkpoint file
model = init_segmentor(config_file, checkpoint_file, device='cuda:0')
tot_frames = 0
video = mmcv.VideoReader(args.video_path)
cap = None
fps = None
if save_out_video:
cap, fps, videoWriter = init_video(args)
det_model = init_detector(
args.det_config, args.det_checkpoint, device=args.device.lower())
# build the pose model from a config file and a checkpoint file
pose_model = init_pose_model(
args.pose_config, args.pose_checkpoint, device=args.device.lower())
dataset = pose_model.cfg.data['test']['type']
for frame in video:
tot_frames += 1
if tot_frames % (fps / 2) == 0:
img = cv2.imread(fpath, cv2.IMREAD_COLOR)
mmdet_results = inference_detector(det_model, img)
person_results = process_mmdet_results(mmdet_results,
args.det_cat_id)
# test a single image, with a list of bboxes.
pose_results, returned_outputs = inference_top_down_pose_model(
pose_model,
img,
person_results,
bbox_thr=args.bbox_thr,
format='xyxy',
dataset=dataset,
return_heatmap=return_heatmap,
outputs=output_layer_names)
result = inference_segmentor(model, frame)
fr_outfile = os.path.join(
args.out_video_root,
pathlib.Path(args.video_path).stem + ".jpg")
fr_resfile = os.path.join(
args.out_video_root,
pathlib.Path(args.video_path).stem + ".json")
vis_img = model.show_result(
frame, result, opacity=0.5, out_file=f'{fr_outfile}')
with open(fr_resfile, 'w', encoding='utf8') as json_file:
json.dump(result, json_file, ensure_ascii=False)
videoWriter.write(vis_img)
if tot_frames % 100 == 0:
print(datetime.now(), tot_frames)
cap.release()
if save_out_video:
videoWriter.release()
if __name__ == '__main__':
main()