-
Notifications
You must be signed in to change notification settings - Fork 9
/
LiviaNET.py
137 lines (116 loc) · 5.16 KB
/
LiviaNET.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
from Blocks import *
import torch.nn.init as init
import torch.nn.functional as F
import pdb
import math
#from layers import *
def croppCenter(tensorToCrop,finalShape):
org_shape = tensorToCrop.shape
diff = org_shape[2] - finalShape[2]
croppBorders = int(diff/2)
return tensorToCrop[:,
:,
croppBorders:org_shape[2]-croppBorders,
croppBorders:org_shape[3]-croppBorders,
croppBorders:org_shape[4]-croppBorders]
def convBlock(nin, nout, kernel_size=3, batchNorm = False, layer=nn.Conv3d, bias=True, dropout_rate = 0.0, dilation = 1):
if batchNorm == False:
return nn.Sequential(
nn.PReLU(),
nn.Dropout(p=dropout_rate),
layer(nin, nout, kernel_size=kernel_size, bias=bias, dilation=dilation)
)
else:
return nn.Sequential(
nn.BatchNorm3d(nin),
nn.PReLU(),
nn.Dropout(p=dropout_rate),
layer(nin, nout, kernel_size=kernel_size, bias=bias, dilation=dilation)
)
def convBatch(nin, nout, kernel_size=3, stride=1, padding=1, bias=False, layer=nn.Conv2d, dilation = 1):
return nn.Sequential(
layer(nin, nout, kernel_size=kernel_size, stride=stride, padding=padding, bias=bias, dilation=dilation),
nn.BatchNorm2d(nout),
#nn.LeakyReLU(0.2)
nn.PReLU()
)
class LiviaNet(nn.Module):
def __init__(self, nClasses):
super(LiviaNet, self).__init__()
# Path-Top
#self.conv1_Top = torch.nn.Conv3d(1, 25, kernel_size=3, stride=1, padding=0, dilation=1, groups=1, bias=True)
self.conv1_Top = convBlock(1, 25)
self.conv2_Top = convBlock(25, 25, batchNorm = True)
self.conv3_Top = convBlock(25, 25, batchNorm = True)
self.conv4_Top = convBlock(25, 50, batchNorm = True)
self.conv5_Top = convBlock(50, 50, batchNorm = True)
self.conv6_Top = convBlock(50, 50, batchNorm = True)
self.conv7_Top = convBlock(50, 75, batchNorm = True)
self.conv8_Top = convBlock(75, 75, batchNorm = True)
self.conv9_Top = convBlock(75, 75, batchNorm = True)
self.fully_1 = nn.Conv3d(150, 400, kernel_size=1)
self.fully_2 = nn.Conv3d(400, 100, kernel_size=1)
self.final = nn.Conv3d(100, nClasses, kernel_size=1)
def forward(self, input):
# get the 3 channels as 5D tensors
y_1 = self.conv1_Top(input[:,0:1,:,:,:])
y_2 = self.conv2_Top(y_1)
y_3 = self.conv3_Top(y_2)
y_4 = self.conv4_Top(y_3)
y_5 = self.conv5_Top(y_4)
y_6 = self.conv6_Top(y_5)
y_7 = self.conv7_Top(y_6)
y_8 = self.conv8_Top(y_7)
y_9 = self.conv9_Top(y_8)
y_3_cropped = croppCenter(y_3,y_9.shape)
y_6_cropped = croppCenter(y_6,y_9.shape)
y = self.fully_1(torch.cat((y_3_cropped, y_6_cropped, y_9), dim=1))
y = self.fully_2(y)
return self.final(y)
class LiviaSemiDenseNet(nn.Module):
def __init__(self, nClasses):
super(LiviaSemiDenseNet, self).__init__()
# Path-Top
# self.conv1_Top = torch.nn.Conv3d(1, 25, kernel_size=3, stride=1, padding=0, dilation=1, groups=1, bias=True)
self.conv1_Top = convBlock(1, 25)
self.conv2_Top = convBlock(25, 25, batchNorm=True)
self.conv3_Top = convBlock(25, 25, batchNorm=True)
self.conv4_Top = convBlock(25, 50, batchNorm=True)
self.conv5_Top = convBlock(50, 50, batchNorm=True)
self.conv6_Top = convBlock(50, 50, batchNorm=True)
self.conv7_Top = convBlock(50, 75, batchNorm=True)
self.conv8_Top = convBlock(75, 75, batchNorm=True)
self.conv9_Top = convBlock(75, 75, batchNorm=True)
self.fully_1 = nn.Conv3d(450, 400, kernel_size=1)
self.fully_2 = nn.Conv3d(400, 100, kernel_size=1)
self.final = nn.Conv3d(100, nClasses, kernel_size=1)
def forward(self, input):
# get the 3 channels as 5D tensors
y_1 = self.conv1_Top(input[:, 0:1, :, :, :])
y_2 = self.conv2_Top(y_1)
y_3 = self.conv3_Top(y_2)
y_4 = self.conv4_Top(y_3)
y_5 = self.conv5_Top(y_4)
y_6 = self.conv6_Top(y_5)
y_7 = self.conv7_Top(y_6)
y_8 = self.conv8_Top(y_7)
y_9 = self.conv9_Top(y_8)
y_1_cropped = croppCenter(y_1, y_9.shape)
y_2_cropped = croppCenter(y_2, y_9.shape)
y_3_cropped = croppCenter(y_3, y_9.shape)
y_4_cropped = croppCenter(y_4, y_9.shape)
y_5_cropped = croppCenter(y_5, y_9.shape)
y_6_cropped = croppCenter(y_6, y_9.shape)
y_7_cropped = croppCenter(y_7, y_9.shape)
y_8_cropped = croppCenter(y_8, y_9.shape)
y = self.fully_1(torch.cat((y_1_cropped,
y_2_cropped,
y_3_cropped,
y_4_cropped,
y_5_cropped,
y_6_cropped,
y_7_cropped,
y_8_cropped,
y_9), dim=1))
y = self.fully_2(y)
return self.final(y)