main.py
4.09 KB
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
import cv2
import numpy as np
import dlib
import light_remover as lr
import imutils
from imutils import face_utils
from imutils.video import VideoStream
from keras.models import load_model
import pygame
def main():
leye = cv2.CascadeClassifier('pre_trained/haarcascade_lefteye_2splits.xml')
reye = cv2.CascadeClassifier('pre_trained/haarcascade_righteye_2splits.xml')
model = load_model('model/eyes_weight.h5')
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor('pre_trained/shape_predictor_68_face_landmarks.dat')
pygame.mixer.init()
warning_sound = pygame.mixer.Sound('asset/windsheld.mp3')
danger_sound = pygame.mixer.Sound('asset/pullup.mp3')
video = VideoStream(src=0).start()
# const
font = cv2.FONT_HERSHEY_COMPLEX_SMALL
target_size = (24,24)
# variable
status = 0 # 0: stay awake, 1: drowsy, 2: sleep
both_eyes_closed_count = 0
def get_eye_open(frame, bound) -> bool:
for (x,y,w,h) in bound:
eye = frame[y:y+h, x:x+w]
cv2.rectangle(frame, (x,y), (x+w,y+h), (0,255,0), 2) # Paint eye bound box
eye = cv2.cvtColor(eye, cv2.COLOR_BGR2GRAY) # using only for reshaping
eye = cv2.resize(eye, target_size)
eye = eye/255
eye = eye.reshape(24, 24, -1)
eye = np.expand_dims(eye, axis=0)
prob = model.predict(eye, verbose=0)
pred = prob.argmax(axis=-1)
return True if pred[0] == 1 else False
def get_eye_open_dlib(frame, bound):
l, r = min(map(lambda x: x[0], bound)), max(map(lambda x: x[0], bound))
t, b = min(map(lambda x: x[1], bound)), max(map(lambda x: x[1], bound))
cv2.rectangle(frame, (l,t), (r,b), (0,255,0), 2)
while True:
frame = video.read()
frame = imutils.resize(frame, width=640)
height, width = frame.shape[:2]
cv2.rectangle(frame, (0, height-50), (200, height), (255,0,0), thickness=cv2.FILLED)
L, gray = lr.light_removing(frame)
left_eye = leye.detectMultiScale(gray)
right_eye = reye.detectMultiScale(gray)
rects = detector(gray, 0)
left_eye_open = get_eye_open(frame, left_eye)
right_eye_open = get_eye_open(frame, right_eye)
# For each detected face, find the landmark.
for (i, rect) in enumerate(rects):
# Make the prediction and transfom it to numpy array
shape = predictor(gray, rect)
shape = face_utils.shape_to_np(shape)
r_eye = shape[36:42]
l_eye = shape[42:48]
mouth = shape[48:]
get_eye_open_dlib(frame, r_eye)
get_eye_open_dlib(frame, l_eye)
for (x, y) in mouth:
cv2.circle(frame, (x, y), 2, (0, 255, 0), -1)
both_eyes_closed = False
if left_eye_open != None or right_eye_open != None:
both_eyes_closed = not left_eye_open and not right_eye_open
if both_eyes_closed:
both_eyes_closed_count += 1
else:
both_eyes_closed_count = max(both_eyes_closed_count - 1, 0)
cv2.putText(frame, "Closed" if both_eyes_closed else "Open", (10, height-20), font, 1, (255,255,255), 1, cv2.LINE_AA)
cv2.putText(frame, f"Count: {both_eyes_closed_count}", (100, height-20), font, 1, (255,255,255), 1, cv2.LINE_AA)
if both_eyes_closed_count < 15:
if status != 0:
warning_sound.stop()
status = 0
elif both_eyes_closed_count < 50:
if status != 1:
danger_sound.stop()
warning_sound.play(-1)
cv2.rectangle(frame, (0,0), (width,height), (0,0,255), both_eyes_closed_count // 5)
status = 1
else:
if status != 2:
warning_sound.stop()
danger_sound.play(-1)
cv2.rectangle(frame, (0,0), (width,height), (0,0,255), 20)
status = 2
cv2.imshow('frame', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cv2.destroyAllWindows()
video.stop()
if __name__ == '__main__':
main()