-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathSegnet.py
210 lines (170 loc) · 8.9 KB
/
Segnet.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
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
import torch
from torch import nn
from torch.nn import functional as F
import torch as t
class Segnet(nn.Module):
def __init__(self, input_nc, output_nc): # 将 init 改为 __init__
super(Segnet, self).__init__()
# Encoder
self.conv11 = nn.Conv2d(input_nc, 64, kernel_size=3, padding=1) ##[4,256,256]-->[64,256,256]
self.bn11 = nn.BatchNorm2d(64)
self.conv12 = nn.Conv2d(64, 64, kernel_size=3, padding=1)
self.bn12 = nn.BatchNorm2d(64)
self.conv21 = nn.Conv2d(64, 128, kernel_size=3, padding=1)
self.bn21 = nn.BatchNorm2d(128)
self.conv22 = nn.Conv2d(128, 128, kernel_size=3, padding=1)
self.bn22 = nn.BatchNorm2d(128)
self.conv31 = nn.Conv2d(128, 256, kernel_size=3, padding=1)
self.bn31 = nn.BatchNorm2d(256)
self.conv32 = nn.Conv2d(256, 256, kernel_size=3, padding=1)
self.bn32 = nn.BatchNorm2d(256)
self.conv33 = nn.Conv2d(256, 256, kernel_size=3, padding=1)
self.bn33 = nn.BatchNorm2d(256)
self.conv41 = nn.Conv2d(256, 512, kernel_size=3, padding=1)
self.bn41 = nn.BatchNorm2d(512)
self.conv42 = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn42 = nn.BatchNorm2d(512)
self.conv43 = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn43 = nn.BatchNorm2d(512)
self.conv51 = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn51 = nn.BatchNorm2d(512)
self.conv52 = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn52 = nn.BatchNorm2d(512)
self.conv53 = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn53 = nn.BatchNorm2d(512)
# Decoder
self.conv53d = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn53d = nn.BatchNorm2d(512)
self.conv52d = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn52d = nn.BatchNorm2d(512)
self.conv51d = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn51d = nn.BatchNorm2d(512)
self.conv43d = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn43d = nn.BatchNorm2d(512)
self.conv42d = nn.Conv2d(512, 512, kernel_size=3, padding=1)
self.bn42d = nn.BatchNorm2d(512)
self.conv41d = nn.Conv2d(512, 256, kernel_size=3, padding=1)
self.bn41d = nn.BatchNorm2d(256)
self.conv33d = nn.Conv2d(256, 256, kernel_size=3, padding=1)
self.bn33d = nn.BatchNorm2d(256)
self.conv32d = nn.Conv2d(256, 256, kernel_size=3, padding=1)
self.bn32d = nn.BatchNorm2d(256)
self.conv31d = nn.Conv2d(256, 128, kernel_size=3, padding=1)
self.bn31d = nn.BatchNorm2d(128)
self.conv22d = nn.Conv2d(128, 128, kernel_size=3, padding=1)
self.bn22d = nn.BatchNorm2d(128)
self.conv21d = nn.Conv2d(128, 64, kernel_size=3, padding=1)
self.bn21d = nn.BatchNorm2d(64)
self.conv12d = nn.Conv2d(64, 64, kernel_size=3, padding=1)
self.bn12d = nn.BatchNorm2d(64)
self.conv11d = nn.Conv2d(64, output_nc, kernel_size=3, padding=1)
def forward(self, x):
# Stage 1
x11 = F.relu(self.bn11(self.conv11(x)), inplace=True)
x12 = F.relu(self.bn12(self.conv12(x11)), inplace=True)
x1p, id1 = F.max_pool2d(x12, kernel_size=2, stride=2, return_indices=True)
# Stage 2
x21 = F.relu(self.bn21(self.conv21(x1p)), inplace=True)
x22 = F.relu(self.bn22(self.conv22(x21)), inplace=True)
x2p, id2 = F.max_pool2d(x22, kernel_size=2, stride=2, return_indices=True)
# Stage 3
x31 = F.relu(self.bn31(self.conv31(x2p)), inplace=True)
x32 = F.relu(self.bn32(self.conv32(x31)), inplace=True)
x33 = F.relu(self.bn33(self.conv33(x32)), inplace=True)
x3p, id3 = F.max_pool2d(x33, kernel_size=2, stride=2, return_indices=True)
# Stage 4
x41 = F.relu(self.bn41(self.conv41(x3p)), inplace=True)
x42 = F.relu(self.bn42(self.conv42(x41)), inplace=True)
x43 = F.relu(self.bn43(self.conv43(x42)), inplace=True)
x4p, id4 = F.max_pool2d(x43, kernel_size=2, stride=2, return_indices=True)
# Stage 5
x51 = F.relu(self.bn51(self.conv51(x4p)), inplace=True)
x52 = F.relu(self.bn52(self.conv52(x51)), inplace=True)
x53 = F.relu(self.bn53(self.conv53(x52)), inplace=True)
x5p, id5 = F.max_pool2d(x53, kernel_size=2, stride=2, return_indices=True)
# Stage 5d
x5d = F.max_unpool2d(x5p, id5, kernel_size=2, stride=2)
x53d = F.relu(self.bn53d(self.conv53d(x5d)), inplace=True)
x52d = F.relu(self.bn52d(self.conv52d(x53d)), inplace=True)
x51d = F.relu(self.bn51d(self.conv51d(x52d)), inplace=True)
# Stage 4d
x4d = F.max_unpool2d(x51d, id4, kernel_size=2, stride=2)
x43d = F.relu(self.bn43d(self.conv43d(x4d)), inplace=True)
x42d = F.relu(self.bn42d(self.conv42d(x43d)), inplace=True)
x41d = F.relu(self.bn41d(self.conv41d(x42d)), inplace=True)
# Stage 3d
x3d = F.max_unpool2d(x41d, id3, kernel_size=2, stride=2)
x33d = F.relu(self.bn33d(self.conv33d(x3d)), inplace=True)
x32d = F.relu(self.bn32d(self.conv32d(x33d)), inplace=True)
x31d = F.relu(self.bn31d(self.conv31d(x32d)), inplace=True)
# Stage 2d
x2d = F.max_unpool2d(x31d, id2, kernel_size=2, stride=2)
x22d = F.relu(self.bn22d(self.conv22d(x2d)), inplace=True)
x21d = F.relu(self.bn21d(self.conv21d(x22d)), inplace=True)
# Stage 1d
x1d = F.max_unpool2d(x21d, id1, kernel_size=2, stride=2)
x12d = F.relu(self.bn12d(self.conv12d(x1d)), inplace=True)
x11d = self.conv11d(x12d)
# output = t.sigmoid(x11d)
return x11d
class UNet(nn.Module):
def __init__(self, in_channels=3, num_classes=2, base_num_filters=16):
super(UNet, self).__init__()
# Down-sampling path (contracting path)
self.conv1 = nn.Conv2d(in_channels, base_num_filters, kernel_size=3,
padding=1) # (N, 3, 128, 128)->(N, 16, 128, 128)
# max pooling
self.conv2 = nn.Conv2d(base_num_filters, base_num_filters * 2, kernel_size=3,
padding=1) # (N, 16, 64, 64)->(N, 32, 64, 64)
# max pooling
self.conv3 = nn.Conv2d(base_num_filters * 2, base_num_filters * 4, kernel_size=3,
padding=1) # (N, 32, 32, 32)->(N, 128, 32, 32)
# max pooling
self.conv4 = nn.Conv2d(base_num_filters * 4, base_num_filters * 8, kernel_size=3,
padding=1) # (N, 128, 16, 16)->(N, 256, 16, 16)
# max pooling
self.conv5 = nn.Conv2d(base_num_filters * 8, base_num_filters * 16, kernel_size=3,
padding=1) # (N, 256, 8, 8)->(N, 512, 8, 8)
# Up-sampling path (expansive path)
self.upconv4 = nn.ConvTranspose2d(base_num_filters * 16, base_num_filters * 8, kernel_size=2,
stride=2) # (N, 512, 8, 8)->(N, 256, 16, 16)
# concatenate (upconv4, conv4)
self.conv6 = nn.Conv2d(base_num_filters * 16, base_num_filters * 8, kernel_size=3, padding=1)
self.upconv3 = nn.ConvTranspose2d(base_num_filters * 8, base_num_filters * 4, kernel_size=2, stride=2)
# concatenate (upconv3, conv3)
self.conv7 = nn.Conv2d(base_num_filters * 8, base_num_filters * 4, kernel_size=3, padding=1)
self.upconv2 = nn.ConvTranspose2d(base_num_filters * 4, base_num_filters * 2, kernel_size=2, stride=2)
# concatenate (upconv2, conv2)
self.conv8 = nn.Conv2d(base_num_filters * 4, base_num_filters * 2, kernel_size=3, padding=1)
self.upconv1 = nn.ConvTranspose2d(base_num_filters * 2, base_num_filters, kernel_size=2, stride=2)
# concatenate (upconv1, conv1)
self.conv9 = nn.Conv2d(base_num_filters * 2, base_num_filters, kernel_size=3, padding=1)
# Final layer
self.conv10 = nn.Conv2d(base_num_filters, num_classes, kernel_size=1)
def forward(self, x):
# Down-sampling path
c1 = F.relu(self.conv1(x))
p1 = F.max_pool2d(c1, kernel_size=2, stride=2)
c2 = F.relu(self.conv2(p1))
p2 = F.max_pool2d(c2, kernel_size=2, stride=2)
c3 = F.relu(self.conv3(p2))
p3 = F.max_pool2d(c3, kernel_size=2, stride=2)
c4 = F.relu(self.conv4(p3))
p4 = F.max_pool2d(c4, kernel_size=2, stride=2)
c5 = F.relu(self.conv5(p4))
# Up-sampling path
up4 = self.upconv4(c5)
up4 = torch.cat([up4, c4], dim=1)
c6 = F.relu(self.conv6(up4))
up3 = self.upconv3(c6)
up3 = torch.cat([up3, c3], dim=1)
c7 = F.relu(self.conv7(up3))
up2 = self.upconv2(c7)
up2 = torch.cat([up2, c2], dim=1)
c8 = F.relu(self.conv8(up2))
up1 = self.upconv1(c8)
up1 = torch.cat([up1, c1], dim=1)
c9 = F.relu(self.conv9(up1))
# Final layer
out = self.conv10(c9)
return out