-
Notifications
You must be signed in to change notification settings - Fork 2
/
sevimon.py
executable file
·182 lines (147 loc) · 6.15 KB
/
sevimon.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
#!/usr/bin/python3
# Licensed under AGPLv3+
import os
import sys
os.environ["OPENCV_LOG_LEVEL"]="FATAL" # Suppress OpenCV messages
import cv2
import numpy
from datetime import datetime,timedelta
import platformdirs
from hsemotion_onnx.facial_emotions import HSEmotionRecognizer
sys.path.append(os.path.dirname(__file__))
from lib.centerface import CenterFace
from lib.i18n import _
from lib.config import *
from lib.cam import *
def writestat(cfg, i, scores) -> None:
now = datetime.now()
print(f'{now.strftime("%H:%M:%S")} ', end='')
# Print scores
emax = numpy.argmax(scores)
print(f'[{i}]: {emotions[emax]}; ', end='')
for e in range(len(emotions)):
print(f'{emotions[e]}: {scores[e]:4.1f}', end='')
if e < (len(emotions) - 1):
print(f', ', end='')
else:
print('')
# Write to logs
if cfg.writestat:
if not os.path.exists(platformdirs.user_log_dir(ANAME)):
os.makedirs(platformdirs.user_log_dir(ANAME))
fp = open(platformdirs.user_log_dir(ANAME) + "/" + now.strftime("%Y.%m.%d"), 'a')
str = now.strftime("%H:%M:%S")
for e in range(len(emotions)):
str = str + " %4.1lf" % (scores[e])
str = str + "\n"
fp.write(str)
fp.close()
# Parameters: configuration, scores, warning was set before
# waring action was active before, last waring was turned on time
def warn_actions(cfg, scores, wws, wwact, wstime):
# Check warning state and notify user
wname = "Face warn"
ws = False # Warning state flag
for e in range(len(emotions)):
if cfg.wminen[e] and scores[e] < cfg.wmin[e]:
ws = True
break
if cfg.wmaxen[e] and scores[e] > cfg.wmax[e]:
ws = True
break
if wws is False and ws is True: # If warning state is switched on
wstime = datetime.now() # Remember time
# Set/reset warning action depending on warning state and it's timeout
if ws and wstime + timedelta(seconds=cfg.wdelay) <= datetime.now():
wact = True
else:
wact = False
try:
# Show warning window in case of:
# it's warning condition, it's on in cfg, has no warning window yet
if wact and cfg.showwarn and not wwact:
# Use OpenCV to avoid excess dependencies
w = cfg.wsize
h = cfg.wsize
if sys.platform == 'darwin' and w < 200:
# Mac OS doesn't allow windows width less than 200, so...
w = 200
wimg = numpy.zeros((h, w, 3), numpy.uint8)
wimg = cv2.rectangle(wimg, (0, 0), (w - 1, h - 1), cfg.wcolor[::-1], -1)
font = cv2.FONT_HERSHEY_SIMPLEX
text = "!"
linew = int(h / 32)
textsize = cv2.getTextSize(text, font, 1, linew)[0]
textx = int((wimg.shape[1] - textsize[0]) / 2)
texty = int((wimg.shape[0] + textsize[1]) / 2)
cv2.putText(wimg, "!", (textx, texty), font, 1, (0, 0, 0), linew)
cv2.namedWindow(wname, cv2.WINDOW_NORMAL | cv2.WINDOW_FREERATIO | cv2.WINDOW_GUI_NORMAL)
cv2.resizeWindow(wname, w, h)
cv2.moveWindow(wname, cfg.wpos[0], cfg.wpos[1])
cv2.setWindowProperty(wname, cv2.WND_PROP_TOPMOST, 1)
cv2.imshow(wname, wimg)
# Destroy warning window in case of:
# it's not to be shown (cfg and condition) and has previous window
elif not (cfg.showwarn and wact) and wwact:
cv2.destroyWindow(wname)
if wact and cfg.beepwarn:
# Generate system beep
print("\a", end="")
except Exception as exc:
print(f'Warning: {exc}')
return ws, wact, wstime
def main() -> None:
MODEL_NAME='enet_b0_8_best_vgaf'
cfg = readcfg()
cam = cam_class(cfg)
bestface = 0 # If configured, here is index of the most intresting face
wws = False # Warning condition was set
wwact = False # Warning windows was shown flag
wstime = datetime.now() # Warning condition set time
# Set neural networks
centerface = CenterFace()
fer = HSEmotionRecognizer(MODEL_NAME)
ret, cap = cam.find_camera()
if ret is False:
return -1
# Set main window properties
cv2.namedWindow('Video', cv2.WINDOW_GUI_NORMAL | cv2.WINDOW_AUTOSIZE)
while True:
ret, image_bgr = cam.get_next_frame()
if not ret:
print(_("Can't read camera image"))
return -1
image = cv2.cvtColor(image_bgr, cv2.COLOR_BGR2RGB)
bounding_boxes, ign = centerface(image_bgr, image_bgr.shape[0], image_bgr.shape[1], threshold=0.35)
if not cfg.allfaces:
# Find biggest (i.e. closer to camera) face
areas = (bounding_boxes[:, 2] - bounding_boxes[:, 0] + 1) * (bounding_boxes[:, 3] - bounding_boxes[:, 1] + 1)
if areas.size != 0:
bestface = areas.argmax()
for i in range(len(bounding_boxes)):
bbox = bounding_boxes[i]
x1, y1, x2, y2 = [round(b) for b in bbox[0:4]]
if (x1 <= 0): x1 = 0
if (y1 <= 0): y1 = 0
if (x2 >= image_bgr.shape[1]): x2 = image_bgr.shape[1] - 1
if (y2 >= image_bgr.shape[0]): y2 = image_bgr.shape[0] - 1
# Just draw border around not selected face
if not cfg.allfaces and i != bestface:
cv2.rectangle(image_bgr, (x1, y1), (x2, y2), (255, 255, 255), 1)
continue
face_img = image[y1:y2, x1:x2]
emotion, scores = fer.predict_emotions(face_img,logits=True)
cv2.rectangle(image_bgr, (x1, y1), (x2, y2), (255, 0, 0), 1)
wws, wwact, wstime = warn_actions(cfg, scores, wws, wwact, wstime)
writestat(cfg, i, scores)
if cfg.showcap:
# Exit if main window was closed by user
if cv2.getWindowProperty('Video', cv2.WND_PROP_VISIBLE) < 1:
break
# Show captured image with the detected box
cv2.imshow('Video', image_bgr)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
if __name__ == "__main__":
main()
# vi: tabstop=4 shiftwidth=4 expandtab