-
Notifications
You must be signed in to change notification settings - Fork 0
/
ResUNet.py
116 lines (95 loc) · 3.74 KB
/
ResUNet.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
#An implementation of ResUNet in 'Road extraction by deep residual unet'
import torch
import torch.nn as nn
import torchvision
def conv1x1(in_planes, out_planes, stride=1):
"""1x1 convolution"""
return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False)
def conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=dilation, groups=groups, bias=False, dilation=dilation)
class DecBlock(nn.Module):
def __init__(self, in_channels, out_channels):
super().__init__()
self.up = nn.Upsample(scale_factor=2, mode='bilinear')
self.bn1 = nn.BatchNorm2d(in_channels)
self.conv1 = conv3x3(in_channels, out_channels)
self.bn2 = nn.BatchNorm2d(out_channels)
self.conv2 = conv3x3(out_channels, out_channels)
self.relu = nn.ReLU()
self.identity = nn.Sequential(
nn.BatchNorm2d(in_channels),
conv1x1(in_channels, out_channels))
def forward(self, x, low_feat):
x = self.up(x)
x = torch.cat((x, low_feat), dim=1)
identity = self.identity(x)
x = self.bn1(x)
x = self.relu(x)
x = self.conv1(x)
x = self.bn2(x)
x = self.relu(x)
x = self.conv2(x)
return x + identity
class EncBlock(nn.Module):
def __init__(self, in_channels, out_channels, stride=2):
super().__init__()
self.bn1 = nn.BatchNorm2d(in_channels)
self.conv1 = conv3x3(in_channels, out_channels, stride)
self.bn2 = nn.BatchNorm2d(out_channels)
self.conv2 = conv3x3(out_channels, out_channels)
self.relu = nn.ReLU()
self.downsample = nn.Sequential(
nn.BatchNorm2d(in_channels),
conv1x1(in_channels, out_channels, stride))
def forward(self, x):
identity = self.downsample(x)
x = self.bn1(x)
x = self.relu(x)
x = self.conv1(x)
x = self.bn2(x)
x = self.relu(x)
x = self.conv2(x)
return x + identity
class Enc0(nn.Module):
def __init__(self, in_channels, out_channels, stride=1):
super().__init__()
self.conv1 = conv3x3(in_channels, out_channels, stride)
self.bn1 = nn.BatchNorm2d(out_channels)
self.conv2 = conv3x3(out_channels, out_channels)
self.relu = nn.ReLU()
self.identity = nn.Sequential(
conv1x1(in_channels, out_channels, stride),
nn.BatchNorm2d(out_channels))
def forward(self, x):
identity = self.identity(x)
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.conv2(x)
return x + identity
class ResUNet(nn.Module):
def __init__(self, in_channels=3, num_classes=1, pretrained=True):
super(ResUNet, self).__init__()
self.Enc0 = Enc0(in_channels, 64)
self.Enc1 = EncBlock(64,128)
self.Enc2 = EncBlock(128,256)
self.bridge = EncBlock(256, 512)
self.Dec2 = DecBlock(512+256, 256)
self.Dec1 = DecBlock(256+128, 128)
self.Dec0 = DecBlock(128+64, 64)
self.classifier = nn.Sequential(
nn.BatchNorm2d(64), nn.ReLU(),
nn.Conv2d(64, num_classes, kernel_size=1))
def forward(self, x):
x_size = x.size()
enc0 = self.Enc0(x)
enc1 = self.Enc1(enc0)
enc2 = self.Enc2(enc1)
enc3 = self.bridge(enc2)
dec2 = self.Dec2(enc3, enc2)
dec1 = self.Dec1(dec2, enc1)
dec0 = self.Dec0(dec1, enc0)
out = self.classifier(dec0)
return out