-
Notifications
You must be signed in to change notification settings - Fork 14
/
model.py
175 lines (144 loc) · 6.45 KB
/
model.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
import torch
import torch.nn as nn
class ChannelAttentionModule(nn.Module):
def __init__(self, channel, ratio=16):
super(ChannelAttentionModule, self).__init__()
#使用自适应池化缩减map的大小，保持通道不变
self.avg_pool = nn.AdaptiveAvgPool2d(1)
self.max_pool = nn.AdaptiveMaxPool2d(1)
self.shared_MLP = nn.Sequential(
nn.Conv2d(channel, channel // ratio, 1, bias=False),
nn.ReLU(),
nn.Conv2d(channel // ratio, channel, 1, bias=False)
)
self.sigmoid = nn.Sigmoid()
def forward(self, x):
avgout = self.shared_MLP(self.avg_pool(x))
maxout = self.shared_MLP(self.max_pool(x))
return self.sigmoid(avgout + maxout)
class SpatialAttentionModule(nn.Module):
def __init__(self):
super(SpatialAttentionModule, self).__init__()
self.conv2d = nn.Conv2d(in_channels=2, out_channels=1, kernel_size=7, stride=1, padding=3)
self.sigmoid = nn.Sigmoid()
def forward(self, x):
#map尺寸不变，缩减通道
avgout = torch.mean(x, dim=1, keepdim=True)
maxout, _ = torch.max(x, dim=1, keepdim=True)
out = torch.cat([avgout, maxout], dim=1)
out = self.sigmoid(self.conv2d(out))
return out
class CBAM(nn.Module):
def __init__(self, channel):
super(CBAM, self).__init__()
self.channel_attention = ChannelAttentionModule(channel)
self.spatial_attention = SpatialAttentionModule()
def forward(self, x):
out = self.channel_attention(x) * x
# out = self.spatial_attention(out) * out
return out
class Conv2DBlock(nn.Module):
""" Conv + ReLU + BN"""
def __init__(self, in_dim, out_dim, kernel_size, padding='same', bias=True, **kwargs):
super(Conv2DBlock, self).__init__(**kwargs)
self.conv = nn.Conv2d(in_dim, out_dim, kernel_size=kernel_size, padding=padding, bias=bias)
self.bn = nn.BatchNorm2d(out_dim)
self.relu = nn.ReLU()
def forward(self, x):
x = self.conv(x)
x = self.bn(x)
x = self.relu(x)
return x
class Double2DConv(nn.Module):
""" Conv2DBlock x 2"""
def __init__(self, in_dim, out_dim):
super(Double2DConv, self).__init__()
self.conv_1 = Conv2DBlock(in_dim, out_dim, (3, 3))
self.conv_2 = Conv2DBlock(out_dim, out_dim, (3, 3))
def forward(self, x):
x = self.conv_1(x)
x = self.conv_2(x)
return x
class Double2DConv2(nn.Module):
""" Conv2DBlock x 2"""
def __init__(self, in_dim, out_dim):
super(Double2DConv2, self).__init__()
self.conv_1 = Conv2DBlock(in_dim, out_dim, (1, 1))
self.conv_2 = Conv2DBlock(out_dim, out_dim, (3, 3))
self.conv_3 = Conv2DBlock(in_dim, out_dim, (3, 3))
self.conv_4 = Conv2DBlock(out_dim, out_dim, (3, 3))
self.conv_5 = Conv2DBlock(in_dim, out_dim, (5, 5))
self.conv_6 = Conv2DBlock(out_dim, out_dim, (3, 3))
self.conv_7 = Conv2DBlock(out_dim*3, out_dim, (3, 3))
def forward(self, x):
x1 = self.conv_1(x)
x1 = self.conv_2(x1)
x2 = self.conv_3(x)
x2 = self.conv_4(x2)
x3 = self.conv_5(x)
x3 = self.conv_6(x3)
x = torch.cat([x1, x2, x3], dim=1)
x = self.conv_7(x)
x = x + x2
return x
class Triple2DConv(nn.Module):
def __init__(self, in_dim, out_dim):
super(Triple2DConv, self).__init__()
self.conv_1 = Conv2DBlock(in_dim, out_dim, (3, 3))
self.conv_2 = Conv2DBlock(out_dim, out_dim, (3, 3))
self.conv_3 = Conv2DBlock(out_dim, out_dim, (3, 3))
def forward(self, x):
x = self.conv_1(x)
x = self.conv_2(x)
x = self.conv_3(x)
return x
class TrackNetV2(nn.Module):
""" Original structure but less two layers
Total params: 10,161,411
Trainable params: 10,153,859
Non-trainable params: 7,552
"""
def __init__(self, in_dim=9, out_dim=3):
super(TrackNetV2, self).__init__()
self.down_block_1 = Double2DConv2(in_dim=in_dim, out_dim=64)
self.down_block_2 = Double2DConv2(in_dim=64, out_dim=128)
self.down_block_3 = Double2DConv2(in_dim=128, out_dim=256)
self.bottleneck = Triple2DConv(in_dim=256, out_dim=512)
self.up_block_1 = Double2DConv(in_dim=768, out_dim=256)
self.up_block_2 = Double2DConv(in_dim=384, out_dim=128)
self.up_block_3 = Double2DConv(in_dim=192, out_dim=64)
self.predictor = nn.Conv2d(64, out_dim, (1, 1))
self.sigmoid = nn.Sigmoid()
self.cbam1 = CBAM(channel=256) #only channel attention
self.cbam2 = CBAM(channel=128)
self.cbam3 = CBAM(channel=64)
self.cbam0_2 = CBAM(channel=256)
self.cbam1_2 = CBAM(channel=128)
self.cbam2_2 = CBAM(channel=64)
def forward(self, x):
""" model input shape: (F*3, 288, 512), output shape: (F, 288, 512) """
x1 = self.down_block_1(x) # (64, 288, 512)
x = nn.MaxPool2d((2, 2), stride=(2, 2))(x1) # (64, 144, 256)
x2 = self.down_block_2(x) # (128, 144, 256)
x = nn.MaxPool2d((2, 2), stride=(2, 2))(x2) # (128, 72, 128)
x3 = self.down_block_3(x) # (256, 72, 128), one less conv layer
x = nn.MaxPool2d((2, 2), stride=(2, 2))(x3) # (256, 36, 64)
x = self.bottleneck(x) # (512, 36, 64)
x3 = self.cbam0_2(x3)
x = torch.cat([nn.Upsample(scale_factor=2)(x), x3], dim=1) # (768, 72, 128) 256+512
x = self.up_block_1(x) # (256, 72, 128), one less conv layer
x = self.cbam1(x)
x2 = self.cbam1_2(x2)
x = torch.cat([nn.Upsample(scale_factor=2)(x), x2], dim=1) # (384, 144, 256) 256+128
x = self.up_block_2(x) # (128, 144, 256)
x = self.cbam2(x)
x1 = self.cbam2_2(x1)
x = torch.cat([nn.Upsample(scale_factor=2)(x), x1], dim=1) # (192, 288, 512) 128+64
x = self.up_block_3(x) # (64, 288, 512)
x = self.cbam3(x)
x = self.predictor(x) # (3, 288, 512)
x = self.sigmoid(x)
return x
# from torchsummary import summary
# Tr = TrackNetV2().cuda()
# summary(Tr, (9, 288, 512))