-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathbrush_ink.py
165 lines (128 loc) · 6.35 KB
/
brush_ink.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
import torch
import torch.nn as nn
import config
import math
import torch.nn.functional as F
class HED(nn.Module):
def __init__(self):
super(HED, self).__init__()
self.netVggOne = nn.Sequential(
nn.Conv2d(in_channels=3, out_channels=64, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False)
)
self.netVggTwo = nn.Sequential(
nn.MaxPool2d(kernel_size=2, stride=2),
nn.Conv2d(in_channels=64, out_channels=128, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=128, out_channels=128, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False)
)
self.netVggThr = nn.Sequential(
nn.MaxPool2d(kernel_size=2, stride=2),
nn.Conv2d(in_channels=128, out_channels=256, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=256, out_channels=256, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=256, out_channels=256, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False)
)
self.netVggFou = nn.Sequential(
nn.MaxPool2d(kernel_size=2, stride=2),
nn.Conv2d(in_channels=256, out_channels=512, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=512, out_channels=512, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=512, out_channels=512, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False)
)
self.netVggFiv = nn.Sequential(
nn.MaxPool2d(kernel_size=2, stride=2),
nn.Conv2d(in_channels=512, out_channels=512, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=512, out_channels=512, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False),
nn.Conv2d(in_channels=512, out_channels=512, kernel_size=3, stride=1, padding=1),
nn.ReLU(inplace=False)
)
self.netScoreOne = nn.Conv2d(in_channels=64, out_channels=1, kernel_size=1, stride=1, padding=0)
self.netScoreTwo = nn.Conv2d(in_channels=128, out_channels=1, kernel_size=1, stride=1, padding=0)
self.netScoreThr = nn.Conv2d(in_channels=256, out_channels=1, kernel_size=1, stride=1, padding=0)
self.netScoreFou = nn.Conv2d(in_channels=512, out_channels=1, kernel_size=1, stride=1, padding=0)
self.netScoreFiv = nn.Conv2d(in_channels=512, out_channels=1, kernel_size=1, stride=1, padding=0)
self.netCombine = nn.Sequential(
nn.Conv2d(in_channels=5, out_channels=1, kernel_size=1, stride=1, padding=0),
nn.Sigmoid()
)
self.load_state_dict({strKey.replace('module', 'net'): tenWeight for strKey, tenWeight in torch.load(config.HED_MODEL_DIR).items()})
# end
def forward(self, tenInput):
tenBlue = (tenInput[:, 0:1, :, :] * 255.0) - 104.00698793
tenGreen = (tenInput[:, 1:2, :, :] * 255.0) - 116.66876762
tenRed = (tenInput[:, 2:3, :, :] * 255.0) - 122.67891434
tenInput = torch.cat([ tenBlue, tenGreen, tenRed ], 1)
tenVggOne = self.netVggOne(tenInput)
tenVggTwo = self.netVggTwo(tenVggOne)
tenVggThr = self.netVggThr(tenVggTwo)
tenVggFou = self.netVggFou(tenVggThr)
tenVggFiv = self.netVggFiv(tenVggFou)
tenScoreOne = self.netScoreOne(tenVggOne)
tenScoreTwo = self.netScoreTwo(tenVggTwo)
tenScoreThr = self.netScoreThr(tenVggThr)
tenScoreFou = self.netScoreFou(tenVggFou)
tenScoreFiv = self.netScoreFiv(tenVggFiv)
tenScoreOne = nn.functional.interpolate(input=tenScoreOne, size=(tenInput.shape[2], tenInput.shape[3]), mode='bilinear', align_corners=False)
tenScoreTwo = nn.functional.interpolate(input=tenScoreTwo, size=(tenInput.shape[2], tenInput.shape[3]), mode='bilinear', align_corners=False)
tenScoreThr = nn.functional.interpolate(input=tenScoreThr, size=(tenInput.shape[2], tenInput.shape[3]), mode='bilinear', align_corners=False)
tenScoreFou = nn.functional.interpolate(input=tenScoreFou, size=(tenInput.shape[2], tenInput.shape[3]), mode='bilinear', align_corners=False)
tenScoreFiv = nn.functional.interpolate(input=tenScoreFiv, size=(tenInput.shape[2], tenInput.shape[3]), mode='bilinear', align_corners=False)
return self.netCombine(torch.cat([ tenScoreOne, tenScoreTwo, tenScoreThr, tenScoreFou, tenScoreFiv ], 1))
def no_sigmoid_cross_entropy(sig_logits, label):
count_neg = torch.sum(1.-label)
count_pos = torch.sum(label)
beta = count_neg / (count_pos+count_neg)
pos_weight = beta / (1-beta)
cost = pos_weight * label * (-1) * torch.log(sig_logits) + (1-label)* (-1) * torch.log(1-sig_logits)
cost = torch.mean(cost * (1-beta))
return cost
def gauss_kernel(image, kernel_size=21, sigma=3, channels=3):
# Create a x, y coordinate grid of shape (kernel_size, kernel_size, 2)
x_cord = torch.arange(kernel_size)
x_grid = x_cord.repeat(kernel_size).view(kernel_size, kernel_size)
y_grid = x_grid.t()
xy_grid = torch.stack([x_grid, y_grid], dim=-1)
mean = (kernel_size - 1)/2.
variance = sigma**2.
# Calculate the 2-dimensional gaussian kernel which is
# the product of two gaussian distributions for two different
# variables (in this case called x and y)
gaussian_kernel = (1./(2.*math.pi*variance)) *\
torch.exp(
-torch.sum((xy_grid - mean)**2., dim=-1) /\
(2*variance)
).cuda()
# Make sure sum of values in gaussian kernel equals 1.
gaussian_kernel = gaussian_kernel / torch.sum(gaussian_kernel)
# Reshape to 2d depthwise convolutional weight
gaussian_kernel = gaussian_kernel.view(1, 1, kernel_size, kernel_size)
gaussian_kernel = gaussian_kernel.repeat(channels, 1, 1, 1)
gaussian_filter = nn.Conv2d(in_channels=channels, out_channels=channels,
kernel_size=kernel_size, groups=channels, bias=False)
gaussian_filter.weight.data = gaussian_kernel
gaussian_filter.weight.requires_grad = False
return gaussian_filter(image).cuda()
def erode(image, kernel_size=5):
pad_size = int(kernel_size/2)
p1d = (pad_size,pad_size,pad_size,pad_size)
padding_image = F.pad(image, p1d, "constant", 1)
erode_image = -1 * (F.max_pool2d(-1 * padding_image, kernel_size, 1))
return erode_image
def test():
img_channels = 3
img_size = 256
x = torch.randn((1, img_channels, img_size+200, img_size))
hed = HED()
print(hed(x).shape)
if __name__ == "__main__":
test()