-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathUNetOG.py
163 lines (117 loc) · 4.43 KB
/
UNetOG.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
import torch
import torch.nn as nn
from torchvision.transforms import RandomCrop
from GenericBlocks import DoubleConvSame, Encoder, DoubleConv
class UNet(nn.Module):
"""Original UNet from the paper"""
def __init__(self, c_in, c_out):
super(UNet, self).__init__()
self.conv1 = DoubleConvSame(c_in=c_in, c_out=64)
self.pool = nn.MaxPool2d(kernel_size=2, stride=2)
self.enc1 = Encoder(64)
self.enc2 = Encoder(128)
self.enc3 = Encoder(256)
self.enc4 = Encoder(512)
self.conv5 = DoubleConvSame(c_in=512, c_out=1024)
self.up1 = nn.ConvTranspose2d(
in_channels=1024, out_channels=512, kernel_size=2, stride=2
)
self.up2 = nn.ConvTranspose2d(
in_channels=512, out_channels=256, kernel_size=2, stride=2
)
self.up3 = nn.ConvTranspose2d(
in_channels=256, out_channels=128, kernel_size=2, stride=2
)
self.up4 = nn.ConvTranspose2d(
in_channels=128, out_channels=64, kernel_size=2, stride=2
)
self.up_conv1 = DoubleConv(c_in=1024, c_out=512)
self.up_conv2 = DoubleConv(c_in=512, c_out=256)
self.up_conv3 = DoubleConv(c_in=256, c_out=128)
self.up_conv4 = DoubleConv(c_in=128, c_out=64)
self.conv_1x1 = nn.Conv2d(in_channels=64, out_channels=c_out, kernel_size=1)
def crop_tensor(self, up_tensor, target_tensor):
_, _, H, W = up_tensor.shape
x = RandomCrop(size=(H, W))(target_tensor)
return x
def forward(self, x):
"""ENCODER"""
c1 = self.conv1(x)
p1 = self.pool(c1)
c2, p2 = self.enc1(p1)
c3, p3 = self.enc2(p2)
c4, p4 = self.enc3(p3)
"""BOTTLE-NECK"""
c5 = self.conv5(p4)
"""DECODER"""
u1 = self.up1(c5)
crop1 = self.crop_tensor(u1, c4)
cat1 = torch.cat([u1, crop1], dim=1)
uc1 = self.up_conv1(cat1)
u2 = self.up2(uc1)
crop2 = self.crop_tensor(u2, c3)
cat2 = torch.cat([u2, crop2], dim=1)
uc2 = self.up_conv2(cat2)
u3 = self.up3(uc2)
crop3 = self.crop_tensor(u3, c2)
cat3 = torch.cat([u3, crop3], dim=1)
uc3 = self.up_conv3(cat3)
u4 = self.up4(uc3)
crop4 = self.crop_tensor(u4, c1)
cat4 = torch.cat([u4, crop4], dim=1)
uc4 = self.up_conv4(cat4)
outputs = self.conv_1x1(uc4)
return outputs
class UNet_2(nn.Module):
"""UNet with same padding"""
def __init__(self, c_in, c_out):
super(UNet_2, self).__init__()
self.conv1 = DoubleConvSame(c_in=c_in, c_out=64)
self.pool = nn.MaxPool2d(kernel_size=2, stride=2)
self.enc1 = Encoder(64)
self.enc2 = Encoder(128)
self.enc3 = Encoder(256)
self.enc4 = Encoder(512)
self.conv5 = DoubleConvSame(c_in=512, c_out=1024)
self.pool = nn.MaxPool2d(kernel_size=2, stride=2)
self.up1 = nn.ConvTranspose2d(
in_channels=1024, out_channels=512, kernel_size=2, stride=2
)
self.up2 = nn.ConvTranspose2d(
in_channels=512, out_channels=256, kernel_size=2, stride=2
)
self.up3 = nn.ConvTranspose2d(
in_channels=256, out_channels=128, kernel_size=2, stride=2
)
self.up4 = nn.ConvTranspose2d(
in_channels=128, out_channels=64, kernel_size=2, stride=2
)
self.up_conv1 = DoubleConvSame(c_in=1024, c_out=512)
self.up_conv2 = DoubleConvSame(c_in=512, c_out=256)
self.up_conv3 = DoubleConvSame(c_in=256, c_out=128)
self.up_conv4 = DoubleConvSame(c_in=128, c_out=64)
self.conv_1x1 = nn.Conv2d(in_channels=64, out_channels=c_out, kernel_size=1)
def forward(self, x):
"""ENCODER"""
c1 = self.conv1(x)
p1 = self.pool(c1)
c2, p2 = self.enc1(p1)
c3, p3 = self.enc2(p2)
c4, p4 = self.enc3(p3)
"""BOTTLE-NECK"""
c5 = self.conv5(p4)
"""DECODER"""
u1 = self.up1(c5)
cat1 = torch.cat([u1, c4], dim=1)
uc1 = self.up_conv1(cat1)
u2 = self.up2(uc1)
cat2 = torch.cat([u2, c3], dim=1)
uc2 = self.up_conv2(cat2)
u3 = self.up3(uc2)
cat3 = torch.cat([u3, c2], dim=1)
uc3 = self.up_conv3(cat3)
u4 = self.up4(uc3)
cat4 = torch.cat([u4, c1], dim=1)
uc4 = self.up_conv4(cat4)
outputs = self.conv_1x1(uc4)
return outputs