-
Notifications
You must be signed in to change notification settings - Fork 1
/
lipstick.py
155 lines (132 loc) · 4.46 KB
/
lipstick.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
import sys
import os
import dlib
import glob
import cv2
from skimage import io
from skimage import color
from PIL import Image
from PIL import ImageEnhance
import matplotlib.pyplot as plt
import numpy as np
from scipy.interpolate import interp1d
from pylab import *
r, g, b = (159,19,20.)
up_left_end = 4
up_right_end = 7
in_left_end = 3
in_right_end = 7
lower_left_end = 5
upper_left_end = 11
lower_right_end = 16
upper_right_end = 22
inten = 0.8
def inter(lx, ly, k1='quadratic'):
unew = np.arange(lx[0], lx[-1] + 1, 1)
f2 = interp1d(lx, ly, kind=k1)
return f2, unew
def getpoint(img):
predictor_path = "shape_predictor_68_face_landmarks.dat"
points = []
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor(predictor_path)
x = []
y = []
dets = detector(img, 1)
for k, d in enumerate(dets):
shape = predictor(img, d)
print("Part 0: {}, Part 1: {} ...".format(shape.part(0),
shape.part(1)))
i = 0
for pt in shape.parts():
i = i + 1
x.append(pt.x)
y.append(pt.y)
for i in range(0, len(x) - 1):
if i >= 48 and i <= 59:
pos = (x[i], y[i])
points.append(pos)
pos = (x[48], y[48])
points.append(pos)
for i in range(0, len(x) - 1):
if i >= 60 and i <= 67:
pos = (x[i], y[i])
points.append(pos)
if i == 64:
pos = (x[54], y[54])
points.append(pos)
pos = (x[67], y[67])
points.append(pos)
return points
img = np.array(imread("su.jpg"))
points=getpoint(img)
points=np.array(points)
point_out_x = np.array((points[:12][:, 0]))
point_out_y = np.array(points[:12][:, 1])
point_in_x = np.array(points[12:][:, 0])
point_in_y = np.array(points[12:][:, 1])
im = img.copy()
im2 = img.copy()
o_l = inter([point_out_x[0]] + point_out_x[up_right_end - 1:][::-1].tolist(),
[point_out_y[0]] + point_out_y[up_right_end - 1:][::-1].tolist(), 'quadratic')
o_u = inter( point_out_x[:up_right_end][::-1].tolist(),
point_out_y[:up_right_end][::-1].tolist(), 'quadratic')
i_u = inter( point_in_x[:in_right_end][::-1].tolist(),
point_in_y[:in_right_end][::-1].tolist(), 'quadratic')
i_l = inter([point_in_x[0]] + point_in_x[in_right_end - 1:][::-1].tolist(),
[point_in_y[0]] + point_in_y[in_right_end - 1:][::-1].tolist(), 'quadratic')
x = [] # will contain the x coordinates of points on lips
y = [] # will contain the y coordinates of points on lips
for i in range(int(point_in_x[0]),int(point_in_x[6])):
# k = k + 1
for j in range(int(o_u[0](i)),int(i_u[0](i))):
x.append(j)
y.append(i)
for j in range(int(i_l[0](i)), int(o_l[0](i))):
x.append(j)
y.append(i)
for i in range(int(point_out_x[0]),int(point_in_x[0])):
for j in range(int(o_u[0](i)),int(o_l[0](i))):
x.append(j)
y.append(i)
for i in range(int(point_in_x[6]),int(point_out_x[6])):
for j in range(int(o_u[0](i)),int(o_l[0](i))):
x.append(j)
y.append(i)
#Change RGB color
val = color.rgb2lab((im[x, y] / 255.).reshape(len(x), 1, 3)).reshape(len(x), 3)
L, A, B = mean(val[:, 0]), mean(val[:, 1]), mean(val[:, 2])
L1, A1, B1 = color.rgb2lab(np.array((r / 255., g / 255., b / 255.)).reshape(1, 1, 3)).reshape(3, )
ll, aa, bb = L1 - L, A1 - A, B1 - B
val[:, 0] +=ll*inten
val[:, 1] +=aa*inten
val[:, 2] += bb*inten
#Change HSV, making it more natural,optional
im[x, y] = color.lab2rgb(val.reshape(len(x), 1, 3)).reshape(len(x), 3) * 255
hsv_val = cv2.cvtColor(im,cv2.COLOR_BGR2HSV)
hsv_val2=hsv_val[x,y]
L3,A3,B3=mean(hsv_val2[:, 0]), mean(hsv_val2[:, 1]),mean(hsv_val2[:, 2])
hsv_val2[:,0]-=0
hsv_val2[:,1]-=0
hsv_val2[:,2]+=0
#水润:
hsv_val[x,y]=hsv_val2
im=cv2.cvtColor(hsv_val,cv2.COLOR_HSV2BGR)
for i in range(0,len(x)-1):
k=x[i]
f=y[i]
if(im[k,f][0]==225):
im[k,f]=(220,220,220.)
#guassian blur
height,width = im.shape[:2]
filter = np.zeros((height,width))
cv2.fillConvexPoly(filter,np.array(c_[y, x],dtype = 'int32'),1)
filter = cv2.GaussianBlur(filter,(31,31),0)
# Erosion to reduce blur size
kernel = np.ones((10,10),np.uint8)
filter = cv2.erode(filter,kernel,iterations = 1)
alpha=np.zeros([height,width,3],dtype='float64')
alpha[:,:,0]=filter
alpha[:,:,1]=filter
alpha[:,:,2]=filter
imsave('gaosiblur3131.jpg',(alpha*im+(1-alpha)*im2).astype('uint8'))