-
Notifications
You must be signed in to change notification settings - Fork 8
/
FaceShapeKMeans.py
181 lines (143 loc) · 6.02 KB
/
FaceShapeKMeans.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
import numpy as np
import cv2
import dlib
from sklearn.cluster import KMeans
# load the image
imagepath = "Resources/i7.jpg"
face_cascade_path = "D:\Open CV CLS\opencv-master\data\haarcascades\haarcascade_frontalface_default.xml"
predictor_path = "shape_predictor_68_face_landmarks.dat"
faceCascade = cv2.CascadeClassifier(face_cascade_path)
predictor = dlib.shape_predictor(predictor_path)
image = cv2.imread(imagepath)
image = cv2.resize(image, (500, 500))
original = image.copy()
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
gauss = cv2.GaussianBlur(gray, (3, 3), 0)
faces = faceCascade.detectMultiScale(
gauss,
scaleFactor=1.05,
minNeighbors=5,
minSize=(100, 100),
flags=cv2.CASCADE_SCALE_IMAGE
)
print("found {0} faces!".format(len(faces)))
for (x, y, w, h) in faces:
cv2.rectangle(image, (x, y), (x + w, y + h), (0, 255, 0), 2)
dlib_rect = dlib.rectangle(int(x), int(y), int(x + w), int(y + h))
detected_landmarks = predictor(image, dlib_rect).parts()
landmarks = np.matrix([[p.x, p.y] for p in detected_landmarks])
print(landmarks)
landmark = image.copy()
for idx, point in enumerate(landmarks):
pos = (point[0,0], point[0,1] )
cv2.putText(landmark,str(idx),pos,fontFace=cv2.FONT_HERSHEY_SIMPLEX,fontScale=0.4,color=(0,0,255) )
cv2.circle(landmark, pos, 3, color=(0,255,255))
cv2.imshow("Landmarks by DLib", landmark)
results = original.copy()
for (x, y, w, h) in faces:
cv2.rectangle(results, (x, y), (x + w, y + h), (0, 255, 0), 2)
temp = original.copy()
forehead = temp[y:y + int(0.25 * h), x:x + w]
rows, cols, bands = forehead.shape
X = forehead.reshape(rows * cols, bands)
kmeans = KMeans(n_clusters=2, init='k-means++', max_iter=300, n_init=10, random_state=0)
y_kmeans = kmeans.fit_predict(X)
for i in range(0, rows):
for j in range(0, cols):
if y_kmeans[i * cols + j] == True:
forehead[i][j] = [255, 255, 255]
if y_kmeans[i * cols + j] == False:
forehead[i][j] = [0, 0, 0]
forehead_mid = [int(cols / 2), int(rows / 2)] # midpoint of forehead
lef = 0
pixel_value = forehead[forehead_mid[1], forehead_mid[0]]
for i in range(0, cols):
if forehead[forehead_mid[1], forehead_mid[0] - i].all() != pixel_value.all():
lef = forehead_mid[0] - i
break;
left = [lef, forehead_mid[1]]
rig = 0
for i in range(0, cols):
if forehead[forehead_mid[1], forehead_mid[0] + i].all() != pixel_value.all():
rig = forehead_mid[0] + i
break;
right = [rig, forehead_mid[1]]
print(left, "left")
print(right,"right")
line1 = np.subtract(right + y, left + x)[0]
print(right + y,"right + y")
print(left + x,"left + x")
cv2.line(results, tuple(x + left), tuple(y + right), color=(0, 255, 0), thickness=2)
cv2.putText(results, ' Line 1', tuple(x + left), fontFace=cv2.FONT_HERSHEY_SIMPLEX, fontScale=1, color=(0, 255, 0),
thickness=2)
cv2.circle(results, tuple(x + left), 5, color=(255, 0, 0), thickness=-1)
cv2.circle(results, tuple(y + right), 5, color=(255, 0, 0), thickness=-1)
linepointleft = (landmarks[1, 0], landmarks[1, 1])
linepointright = (landmarks[15, 0], landmarks[15, 1])
line2 = np.subtract(linepointright, linepointleft)[0]
print(line2,"line2")
cv2.line(results, linepointleft, linepointright, color=(0, 255, 0), thickness=2)
cv2.putText(results, ' Line 2', linepointleft, fontFace=cv2.FONT_HERSHEY_SIMPLEX, fontScale=1, color=(0, 255, 0),
thickness=2)
cv2.circle(results, linepointleft, 5, color=(255, 0, 0), thickness=-1)
cv2.circle(results, linepointright, 5, color=(255, 0, 0), thickness=-1)
linepointleft = (landmarks[3, 0], landmarks[3, 1])
linepointright = (landmarks[13, 0], landmarks[13, 1])
line3 = np.subtract(linepointright, linepointleft)[0]
cv2.line(results, linepointleft, linepointright, color=(0, 255, 0), thickness=2)
cv2.putText(results, ' Line 3', linepointleft, fontFace=cv2.FONT_HERSHEY_SIMPLEX, fontScale=1, color=(0, 255, 0),
thickness=2)
cv2.circle(results, linepointleft, 5, color=(255, 0, 0), thickness=-1)
cv2.circle(results, linepointright, 5, color=(255, 0, 0), thickness=-1)
linepointbottom = (landmarks[8, 0], landmarks[8, 1])
linepointtop = (landmarks[8, 0], y)
line4 = np.subtract(linepointbottom, linepointtop)[1]
cv2.line(results, linepointtop, linepointbottom, color=(0, 255, 0), thickness=2)
cv2.putText(results, ' Line 4', linepointbottom, fontFace=cv2.FONT_HERSHEY_SIMPLEX, fontScale=1, color=(0, 255, 0),
thickness=2)
cv2.circle(results, linepointtop, 5, color=(255, 0, 0), thickness=-1)
cv2.circle(results, linepointbottom, 5, color=(255, 0, 0), thickness=-1)
print(line1, line2, line3, line4)
similarity = np.std([line1, line2, line3])
print("similarity=", similarity)
# we use arcustangens for angle calculation
ax, ay = landmarks[3, 0], landmarks[3, 1]
print('ax, ay=', ax, ay)
bx, by = landmarks[4, 0], landmarks[4, 1]
cx, cy = landmarks[5, 0], landmarks[5, 1]
dx, dy = landmarks[6, 0], landmarks[6, 1]
import math
from math import degrees
print("cy-ay",cy - ay)
print("cx - ax",cx - ax)
alpha0 = math.atan2(cy - ay, cx - ax)
alpha1 = math.atan2(dy - by, dx - bx)
alpha = alpha1 - alpha0
angle = abs(degrees(alpha))
print('angle=', angle)
angle = 180 - angle
print('alpha1=', alpha)
print('angle=', angle)
for i in range(1):
if similarity < 10:
if angle < 160:
print('squared shape.Jawlines are more angular')
break
else:
print('round shape.Jawlines are not that angular')
break
if line3 > line1:
if angle < 160:
print('triangle shape.Forehead is more wider')
break
if line4 > line2:
if angle < 160:
print('rectangular. face length is largest and jawline are angular ')
break;
else:
print('oblong. face length is largest and jawlines are not angular')
break;
print("Error")
output = np.concatenate((original, results), axis=1)
cv2.imshow('output', output)
cv2.waitKey(0)