-
Notifications
You must be signed in to change notification settings - Fork 0
/
blink-detection.py
191 lines (146 loc) · 6.36 KB
/
blink-detection.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
import cv2, dlib
import numpy as np
from imutils import face_utils
from keras.models import load_model
import os
import time
from pync import Notifier
import threading
IMG_SIZE = (64,56)
B_SIZE = (34, 26)
margin = 95
class_labels = ['center','left', 'right']
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
font_letter = cv2.FONT_HERSHEY_PLAIN
model = load_model('gazev3.1.h5')
model_b = load_model('blinkdetection.h5')
def detect_gaze(eye_img):
pred_l = model.predict(eye_img)
accuracy = int(np.array(pred_l).max() * 100)
gaze = class_labels[np.argmax(pred_l)]
return gaze
def detect_blink(eye_img):
pred_B = model_b.predict(eye_img)
status = pred_B[0][0]
status = status*100
status = round(status,3)
return status
def crop_eye(img, eye_points):
x1, y1 = np.amin(eye_points, axis=0)
x2, y2 = np.amax(eye_points, axis=0)
cx, cy = (x1 + x2) / 2, (y1 + y2) / 2
w = (x2 - x1) * 1.2
h = w * IMG_SIZE[1] / IMG_SIZE[0]
margin_x, margin_y = w / 2, h / 2
min_x, min_y = int(cx - margin_x), int(cy - margin_y)
max_x, max_y = int(cx + margin_x), int(cy + margin_y)
eye_rect = np.rint([min_x, min_y, max_x, max_y]).astype(int)
eye_img = gray[eye_rect[1]:eye_rect[3], eye_rect[0]:eye_rect[2]]
return eye_img, eye_rect
cap = cv2.VideoCapture(0)
frames_to_blink = 6
blinking_frames = 0
timerv = 0
def increment_timerv():
global timerv
# TODO: change the timer to 60 seconds. 10 seconds just for testing purposes
time.sleep(10)
timerv = 1
increment_thread = threading.Thread(target=increment_timerv)
increment_thread.start()
blink_counter = 0
start_time = time.time()
lowblink = 0
while cap.isOpened():
output = np.zeros((900,820,3), dtype="uint8")
ret, img = cap.read()
img = cv2.flip(img,flipCode = 1)
h,w = (112,128)
if not ret:
break
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = detector(gray)
for face in faces:
shapes = predictor(gray, face)
for n in range(36,42):
x= shapes.part(n).x
y = shapes.part(n).y
next_point = n+1
if n==41:
next_point = 36
x2 = shapes.part(next_point).x
y2 = shapes.part(next_point).y
cv2.line(img,(x,y),(x2,y2),(0,69,255),2)
for n in range(42,48):
x= shapes.part(n).x
y = shapes.part(n).y
next_point = n+1
if n==47:
next_point = 42
x2 = shapes.part(next_point).x
y2 = shapes.part(next_point).y
cv2.line(img,(x,y),(x2,y2),(153,0,153),2)
shapes = face_utils.shape_to_np(shapes)
eye_img_l, eye_rect_l = crop_eye(gray, eye_points=shapes[36:42])
eye_img_r, eye_rect_r = crop_eye(gray, eye_points=shapes[42:48])
eye_img_l_view = cv2.resize(eye_img_l, dsize=(128,112))
eye_img_l_view = cv2.cvtColor(eye_img_l_view,cv2.COLOR_BGR2RGB)
eye_img_r_view = cv2.resize(eye_img_r, dsize=(128,112))
eye_img_r_view = cv2.cvtColor(eye_img_r_view, cv2.COLOR_BGR2RGB)
eye_blink_left = cv2.resize(eye_img_l.copy(), B_SIZE)
eye_blink_right = cv2.resize(eye_img_r.copy(), B_SIZE)
eye_blink_left_i = eye_blink_left.reshape((1, B_SIZE[1], B_SIZE[0], 1)).astype(np.float32) / 255.
eye_blink_right_i = eye_blink_right.reshape((1, B_SIZE[1], B_SIZE[0], 1)).astype(np.float32) / 255.
eye_img_l = cv2.resize(eye_img_l, dsize=IMG_SIZE)
eye_input_g = eye_img_l.copy().reshape((1, IMG_SIZE[1], IMG_SIZE[0], 1)).astype(np.float32) / 255.
status_l = detect_blink(eye_blink_left_i)
gaze = detect_gaze(eye_input_g)
if gaze == class_labels[1]:
blinking_frames += 1
if blinking_frames == frames_to_blink:
os.system("beep -f 2000 -l 1500")
elif gaze == class_labels[2]:
blinking_frames += 1
if blinking_frames == frames_to_blink:
os.system("beep -f 2000 -l 1500")
elif status_l < 0.1:
blinking_frames += 1
os.system("beep -f 2000 -l 1500")
else:
blinking_frames = 0
output = cv2.line(output,(400,200), (400,0),(0,255,0),thickness=2)
cv2.putText(output,"LEFT EYE GAZE",(10,180), font_letter,1, (255,255,51),1)
cv2.putText(output,"LEFT EYE OPENING %",(200,180), font_letter,1, (255,255,51),1)
cv2.putText(output,"RIGHT EYE GAZE",(440,180), font_letter,1, (255,255,51),1)
cv2.putText(output,"RIGHT EYE OPENING %",(621,180), font_letter,1, (255,255,51),1)
if status_l < 10 :
blink_counter += 1
cv2.putText(output,"---BLINKING----",(250,300), font_letter,2, (153,153,255),2)
elapsed_time = time.time() - start_time
blinks_per_minute = (blink_counter / elapsed_time) * 60
if blinks_per_minute < 10 and timerv != 0:
lowblink += 1
title = 'Blinkit'
message = 'You are not blinking frequently enough.'
try:
Notifier.notify('Your blink count is too low!', title='Notification', sound='default')
except Exception as e:
print(f"Notification error: {e}")
cv2.putText(output, f"Blinks per Minute: {blinks_per_minute:.2f}", (10, output.shape[0] - 10), font_letter, 1, (255, 255, 255), 1)
cv2.putText(output, f"x: {x}", (10, output.shape[0] - 60), font_letter, 1, (255, 255, 255), 1)
output[0:112, 0:128] = eye_img_l_view
cv2.putText(output, gaze,(30,150), font_letter,2, (0,255,0),2)
output[0:112, margin+w:(margin+w)+w] = eye_img_l_view
cv2.putText(output,(str(status_l)+"%"),((margin+w),150), font_letter,2, (0,0,255),2)
output[0:112, 2*margin+2*w:(2*margin+2*w)+w] = eye_img_r_view
cv2.putText(output, gaze,((2*margin+2*w)+30,150), font_letter,2, (0,0,255),2)
output[0:112, 3*margin+3*w:(3*margin+3*w)+w] = eye_img_r_view
cv2.putText(output, (str(status_l)+"%"),((3*margin+3*w),150), font_letter,2, (0,0,255),2)
img = cv2.resize(img, (640, 480))
output[235+100:715+100, 80:720] = img
cv2.imshow('result',output)
if cv2.waitKey(1) == ord('q') :
break
cap.release()
cv2.destroyAllWindows()