forked from axinc-ai/ailia-models
-
Notifications
You must be signed in to change notification settings - Fork 0
/
pspnet-hair-segmentation.py
170 lines (133 loc) · 4.21 KB
/
pspnet-hair-segmentation.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
import sys
import time
import numpy as np
import cv2
import ailia
# import original modules
sys.path.append('../../util')
from utils import get_base_parser, update_parser # noqa: E402
from image_utils import load_image # noqa: E402
from model_utils import check_and_download_models # noqa: E402
import webcamera_utils # noqa: E402
# ======================
# Parameters
# ======================
WEIGHT_PATH = 'pspnet-hair-segmentation.onnx'
MODEL_PATH = WEIGHT_PATH + '.prototxt'
REMOTE_PATH =\
'https://storage.googleapis.com/ailia-models/pspnet-hair-segmentation/'
IMAGE_PATH = 'test.jpg'
SAVE_IMAGE_PATH = 'output.png'
IMAGE_HEIGHT = 512
IMAGE_WIDTH = 512
# ======================
# Arguemnt Parser Config
# ======================
parser = get_base_parser(
'Real-time hair segmentation model', IMAGE_PATH, SAVE_IMAGE_PATH
)
args = update_parser(parser)
# ======================
# Utils
# ======================
def sigmoid(x):
return 1.0 / (1.0 + np.exp(-x))
def postprocess(src_img, preds_ailia):
pred = sigmoid(preds_ailia)[0][0]
mask = pred >= 0.5
mask_n = np.zeros((IMAGE_HEIGHT, IMAGE_WIDTH, 3))
mask_n[:, :, 0] = 255
mask_n[:, :, 0] *= mask
image_n = cv2.cvtColor(src_img, cv2.COLOR_RGB2BGR)
# discard padded area
h, w, _ = image_n.shape
delta_h = h - IMAGE_HEIGHT
delta_w = w - IMAGE_WIDTH
top = delta_h // 2
bottom = IMAGE_HEIGHT - (delta_h - top)
left = delta_w // 2
right = IMAGE_WIDTH - (delta_w - left)
mask_n = mask_n[top:bottom, left:right, :]
image_n = image_n * 0.5 + mask_n * 0.5
return image_n
# ======================
# Main functions
# ======================
def recognize_from_image():
# prepare input data
input_data = load_image(
args.input,
(IMAGE_HEIGHT, IMAGE_WIDTH),
normalize_type='ImageNet',
gen_input_ailia=True
)
src_img = load_image(
args.input,
(IMAGE_HEIGHT, IMAGE_WIDTH),
normalize_type='None'
)
# net initialize
net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
# inference
print('Start inference...')
if args.benchmark:
print('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
preds_ailia = net.predict(input_data)
end = int(round(time.time() * 1000))
print(f'\tailia processing time {end - start} ms')
else:
preds_ailia = net.predict(input_data)
# postprocessing
res_img = postprocess(src_img, preds_ailia)
cv2.imwrite(args.savepath, res_img)
print('Script finished successfully.')
def recognize_from_video():
# net initialize
net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
capture = webcamera_utils.get_capture(args.video)
# create video writer if savepath is specified as video format
if args.savepath != SAVE_IMAGE_PATH:
print(
'[WARNING] currently, video results cannot be output correctly...'
)
writer = webcamera_utils.get_writer(
args.savepath, IMAGE_HEIGHT, IMAGE_WIDTH
)
else:
writer = None
while(True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
src_img, input_data = webcamera_utils.preprocess_frame(
frame,
IMAGE_HEIGHT,
IMAGE_WIDTH,
normalize_type='ImageNet'
)
src_img = cv2.resize(src_img, (IMAGE_WIDTH, IMAGE_HEIGHT))
src_img = cv2.cvtColor(src_img, cv2.COLOR_BGR2RGB)
preds_ailia = net.predict(input_data)
res_img = postprocess(src_img, preds_ailia)
cv2.imshow('frame', res_img / 255.0)
# # save results
# if writer is not None:
# writer.write(res_img)
capture.release()
cv2.destroyAllWindows()
if writer is not None:
writer.release()
print('Script finished successfully.')
def main():
# model files check and download
check_and_download_models(WEIGHT_PATH, MODEL_PATH, REMOTE_PATH)
if args.video is not None:
# video mode
recognize_from_video()
else:
# image mode
recognize_from_image()
if __name__ == '__main__':
main()