ReadMe.md

How it's works

1. First import the libaray that we need

import cv2
import numpy as np
import dlib
from playsound import playsound
from imutils import face_utils
from scipy.spatial import distance as dist

2. Connect the webcam to the program

cap = cv2.VideoCapture(0)
while True:
    _, frame = cap.read()
    cv2.imshow('Blink detection', frame)
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

cap.release()
cv2.destroyAllWindows()

3. Make the program to target or detect the face

detector = dlib.get_frontal_face_detector() 

4. Insert the model that we need

predictor = dlib.shape_predictor("shape_predictor_68_face_landmarks.dat") 

5. Change the BGR to the grayscale

gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

6. Detect the face in the grayscale image

faces = detector(gray)

7. Convert the coordinates of facial landmark to a numpy

point = predictor(gray, face)
points = face_utils.shape_to_np(point)

8. Calculate the coordinates/indexes for the up and down lip

def calculate_lip(lips):
     dist1 = dist.euclidean(lips[2], lips[6]) 
     dist2 = dist.euclidean(lips[0], lips[4]) 

9. Calculate the ratio of the lip using the coordinates/indexes that have been obtained

EAR = (A + B) / (2.0 * C)`

10. Determine the lip aspect ratio for yawn and the number of frames the person has yawned

counter = 0
lip_LAR = 0.4
lip_per_frame = 30

11. Insert the indexes of the lip

lips = [60,61,62,63,64,65,66,67]

12. Change the indexes into the numpy array

lip_point = points[lips]

13. Calculate the lip aspect ratio

LAR = calculate_lip(lip_point) 

Demo

Clik the picture to see the Video