Mario-using-Eye-gestures/main.py at main · shahsahil264/Mario-using-Eye-gestures · GitHub

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
import cv2
import dlib
from math import hypot
import numpy as np
import os
from pynput.keyboard import Key, Controller
import time

cap = cv2.VideoCapture(0)
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor("/Users/sahil/Desktop/HCI/shape_predictor_68_face_landmarks.dat")
#Downloaded the above file from "https://github.com/italojs/facial-landmarks-recognition/blob/master/shape_predictor_68_face_landmarks.dat"
keyboard = Controller()
BLINK_THRESH = 4.2
THRESHOLD_THRESH = 30

def get_landmarks(keypoint_img):
    landmarks = predictor(gray, face)
    for i in range(68):
        cv2.circle(keypoint_img, (landmarks.part(i).x,
                                  landmarks.part(i).y), 2, 255, -1)
    return landmarks


def get_left_eye_landmarks(landmarks):
    left_eye_left_pt = (landmarks.part(36).x, landmarks.part(36).y)
    left_eye_right_pt = (landmarks.part(39).x, landmarks.part(39).y)
    left_eye_top_mid = (int((landmarks.part(37).x + landmarks.part(38).x) / 2),
                        int((landmarks.part(37).y + landmarks.part(38).y) / 2))
    left_eye_bottom_mid = (int((landmarks.part(
        40).x + landmarks.part(41).x) / 2), int((landmarks.part(40).y + landmarks.part(41).y) / 2))
    return left_eye_left_pt, left_eye_right_pt, left_eye_top_mid, left_eye_bottom_mid


def get_right_eye_landmarks(landmarks):
    right_eye_left_pt = (landmarks.part(42).x, landmarks.part(42).y)
    right_eye_right_pt = (landmarks.part(45).x, landmarks.part(45).y)
    right_eye_top_mid = (int((landmarks.part(43).x + landmarks.part(44).x) / 2),
                         int((landmarks.part(43).y + landmarks.part(44).y) / 2))
    right_eye_bottom_mid = (int((landmarks.part(47).x + landmarks.part(46).x) / 2),
                            int((landmarks.part(47).y + landmarks.part(46).y) / 2))
    return right_eye_left_pt, right_eye_right_pt, right_eye_top_mid, right_eye_bottom_mid


def get_eye_position(eye_left_pt, eye_right_pt, eye_top_mid, eye_bottom_mid):
    hor_line = cv2.line(
        keypoint_img, eye_left_pt, eye_right_pt, (0, 255, 0), 1)
    ver_line = cv2.line(
        keypoint_img, eye_top_mid, eye_bottom_mid, (0, 255, 0), 1)

    hor_line_len = hypot(
        (eye_left_pt[0] - eye_right_pt[0]), (eye_left_pt[1] - eye_right_pt[1]))
    ver_line_len = hypot(
        (eye_top_mid[0] - eye_bottom_mid[0]), (eye_top_mid[1] - eye_bottom_mid[1]))
    return hor_line, ver_line, hor_line_len, ver_line_len


def get_eye_thresh(eye_region):
    min_x, min_y, max_x, max_y = np.min(eye_region[:, 0]), np.min(eye_region[:, 1]), np.max(eye_region[:, 0]), np.max(
        eye_region[:, 1])

    eye = cv2.resize(
        flip[min_y:max_y, min_x:max_x], None, fx=5, fy=5)
    gray_eye = cv2.cvtColor(eye, cv2.COLOR_BGR2GRAY)
    _, thresh_eye = cv2.threshold(
        gray_eye, THRESHOLD_THRESH, 255, cv2.THRESH_BINARY)
    h, w = thresh_eye.shape
    thresh_eye_half = thresh_eye[0:h, 0:int(w / 2)]
    return gray_eye, thresh_eye, thresh_eye_half


# Decide direction where mario will move
def decide_direction(r_nz, l_nz, r, l, avg_open_eye_ratio):
    os.system("clear")
    ctrl = ""
    if r_nz > l_nz:
        ctrl = "Right"
        keyboard.press(Key.right)

        print("CONTROL : " + ctrl)
    else:
        ctrl = "Left"
        keyboard.press(Key.left)

        print("CONTROL : " + ctrl)
    cv2.imshow("EYE INFO RIGHT", r)
    cv2.imshow("EYE INFO LEFT", l)

    print("AVG OPEN EYE RATIO : " + str(round(avg_open_eye_ratio, 3)))
    print("Right non zero : " + str(l_nz))
    print("left non zero : " + str(r_nz))


while 1:
    ret, frame = cap.read()
    flip = cv2.flip(frame, 1)
    keypoint_img = flip.copy()
    gray = cv2.cvtColor(flip, cv2.COLOR_BGR2GRAY)

    faces = detector(gray)
    for face in faces:
        landmarks = get_landmarks(keypoint_img)
        # LEFT EYEBALL
        left_eye_left_pt, left_eye_right_pt, left_eye_top_mid, left_eye_bottom_mid = get_left_eye_landmarks(landmarks)
        left_hor_line, left_ver_line, left_hor_line_len, left_ver_line_len = get_eye_position(left_eye_left_pt,
                                                                                              left_eye_right_pt,
                                                                                              left_eye_top_mid,
                                                                                              left_eye_bottom_mid)

        left_open_eye_ratio = left_hor_line_len / left_ver_line_len
        temp_left = []
        for i in range(36, 42):
            temp_left.append((landmarks.part(i).x, landmarks.part(i).y))
        left_eye_region = np.array(temp_left, np.int32)
        left_gray_eye, left_thresh_eye, left_thresh_eye_half = get_eye_thresh(left_eye_region)

        # RIGHT EYEBALL
        right_eye_left_pt, right_eye_right_pt, right_eye_top_mid, right_eye_bottom_mid = get_right_eye_landmarks(
            landmarks)

        right_hor_line, right_ver_line, right_hor_line_len, right_ver_line_len = get_eye_position(right_eye_left_pt,
                                                                                                  right_eye_right_pt,
                                                                                                  right_eye_top_mid,
                                                                                                  right_eye_bottom_mid)

        right_open_eye_ratio = right_hor_line_len / right_ver_line_len
        temp_right = []
        for i in range(42, 48):
            temp_right.append((landmarks.part(i).x, landmarks.part(i).y))
        right_eye_region = np.array(temp_right, np.int32)
        right_gray_eye, right_thresh_eye, right_thresh_eye_half = get_eye_thresh(right_eye_region)

        # INFO ABOUT THE EYE

        r = np.hstack((right_gray_eye, right_thresh_eye))
        l = np.hstack((left_gray_eye, left_thresh_eye))

        l_nz = cv2.countNonZero(left_thresh_eye_half)
        r_nz = cv2.countNonZero(right_thresh_eye_half)

        avg_open_eye_ratio = (right_open_eye_ratio + left_open_eye_ratio) / 2

        # DISPLAY
        decide_direction(r_nz, l_nz, r, l, avg_open_eye_ratio)

        # Check if eye is blinking
        if avg_open_eye_ratio > BLINK_THRESH:
            keyboard.press(Key.space)
            time.sleep(0.5)
            keyboard.release(Key.space)
            print("STATUS : BLINKING")
        else:
            print("STATUS : NOT BLINKING")

    cv2.imshow("Facial Keypoints", keypoint_img)
    cv2.imshow("Orignal Image", flip)

    if cv2.waitKey(1) & 0xFF == ord("q"):
        break
cap.release()
cv2.destroyAllWindows()