-
Notifications
You must be signed in to change notification settings - Fork 9
/
encoder.py
178 lines (156 loc) · 7.01 KB
/
encoder.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
# -*- coding:utf-8 -*-
# -----------------------------------------
# Filename: encoder.py
# Author : Qing Wu
# Email : [email protected]
# Date : 2021/9/20
# -----------------------------------------
import torch.nn as nn
import torch
# -------------------------------
# RDN encoder network
# <Zhang, Yulun, et al. "Residual dense network for image super-resolution.">
# Here code is modified from: https://github.com/yjn870/RDN-pytorch/blob/master/models.py
# -------------------------------
class DenseLayer(nn.Module):
def __init__(self, in_channels, out_channels):
super(DenseLayer, self).__init__()
self.conv = nn.Conv3d(in_channels, out_channels, kernel_size=3, padding=3 // 2)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
return torch.cat([x, self.relu(self.conv(x))], 1)
class RDB(nn.Module):
def __init__(self, in_channels, growth_rate, num_layers):
super(RDB, self).__init__()
self.layers = nn.Sequential(
*[DenseLayer(in_channels + growth_rate * i, growth_rate) for i in range(num_layers)])
# local feature fusion
self.lff = nn.Conv3d(in_channels + growth_rate * num_layers, growth_rate, kernel_size=1)
def forward(self, x):
return x + self.lff(self.layers(x)) # local residual learning
class RDN(nn.Module):
def __init__(self, feature_dim=128, num_features=64, growth_rate=64, num_blocks=8, num_layers=3):
super(RDN, self).__init__()
self.G0 = num_features
self.G = growth_rate
self.D = num_blocks
self.C = num_layers
# shallow feature extraction
self.sfe1 = nn.Conv3d(1, num_features, kernel_size=3, padding=3 // 2)
self.sfe2 = nn.Conv3d(num_features, num_features, kernel_size=3, padding=3 // 2)
# residual dense blocks
self.rdbs = nn.ModuleList([RDB(self.G0, self.G, self.C)])
for _ in range(self.D - 1):
self.rdbs.append(RDB(self.G, self.G, self.C))
# global feature fusion
self.gff = nn.Sequential(
nn.Conv3d(self.G * self.D, self.G0, kernel_size=1),
nn.Conv3d(self.G0, self.G0, kernel_size=3, padding=3 // 2)
)
self.output = nn.Conv3d(self.G0, feature_dim, kernel_size=3, padding=3 // 2)
def forward(self, x):
sfe1 = self.sfe1(x)
sfe2 = self.sfe2(sfe1)
x = sfe2
local_features = []
for i in range(self.D):
x = self.rdbs[i](x)
local_features.append(x)
x = self.gff(torch.cat(local_features, 1)) + sfe1 # global residual learning
x = self.output(x)
return x
# -------------------------------
# ResCNN encoder network
# <Du, Jinglong, et al. "Super-resolution reconstruction of single
# anisotropic 3D MR images using residual convolutional neural network.">
# -------------------------------
class ResCNN(nn.Module):
def __init__(self, feature_dim=128):
super(ResCNN, self).__init__()
self.conv_start = nn.Sequential(
nn.Conv3d(in_channels=1, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True)
)
self.block1 = nn.Sequential(
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True)
)
self.block2 = nn.Sequential(
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True)
)
self.block3 = nn.Sequential(
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True),
nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, padding=3 // 2),
nn.ReLU(inplace=True)
)
self.conv_end = nn.Sequential(
nn.Conv3d(in_channels=64, out_channels=feature_dim, kernel_size=3, padding=3 // 2),
)
def forward(self, x):
in_block1 = self.conv_start(x)
out_block1 = self.block1(in_block1)
in_block2 = out_block1 + in_block1
out_block2 = self.block2(in_block2)
in_block3 = out_block2 + in_block2
out_block3 = self.block3(in_block3)
res_img = self.conv_end(out_block3 + in_block3)
return x + res_img
# -------------------------------
# SRResNet
# <Ledig, Christian, et al. "Photo-realistic single image super-resolution
# using a generative adversarial network.">
# -------------------------------
def conv(ni, nf, kernel_size=3, actn=False):
layers = [nn.Conv3d(ni, nf, kernel_size, padding=kernel_size // 2)]
if actn: layers.append(nn.ReLU(True))
return nn.Sequential(*layers)
class ResSequential(nn.Module):
def __init__(self, layers, res_scale=1.0):
super().__init__()
self.res_scale = res_scale
self.m = nn.Sequential(*layers)
def forward(self, x): return x + self.m(x) * self.res_scale
def res_block(nf):
return ResSequential(
[conv(nf, nf, actn=True), conv(nf, nf)],
1.0) # this is best one
class SRResnet(nn.Module):
def __init__(self, nf=64, feature_dim=128):
super().__init__()
features = [conv(1, nf)]
for i in range(18): features.append(res_block(nf))
features += [conv(nf, nf),
conv(nf, feature_dim)]
self.features = nn.Sequential(*features)
def forward(self, x):
return self.features(x)