UNet++详细解读（二）pytorch从头开始搭建UNet++

Unet++代码

网络架构

黑色部分是Backbone，是原先的UNet。

绿色箭头为上采样，蓝色箭头为密集跳跃连接。

绿色的模块为密集连接块，是经过左边两个部分拼接操作后组成的

Backbone

2个3x3的卷积，padding=1。

class VGGBlock(nn.Module):
    def __init__(self, in_channels, middle_channels, out_channels):
        super().__init__()
        self.relu = nn.ReLU(inplace=True)
        self.conv1 = nn.Conv2d(in_channels, middle_channels, 3, padding=1)
        self.bn1 = nn.BatchNorm2d(middle_channels)
        self.conv2 = nn.Conv2d(middle_channels, out_channels, 3, padding=1)
        self.bn2 = nn.BatchNorm2d(out_channels)

    def forward(self, x):
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)

        return out
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

上采样

图中的绿色箭头，上采样使用双线性插值。

双线性插值就是有两个变量的插值函数的线性插值扩展，其核心思想是在两个方向分别进行一次线性插值

torch.nn.Upsample(size=None, scale_factor=None, mode='nearest', align_corners=None, recompute_scale_factor=None)
1

参数说明：
①size：可以用来指定输出空间的大小，默认是None；
②scale_factor：比例因子，比如scale_factor=2意味着将输入图像上采样2倍，默认是None；
③mode：用来指定上采样算法，有’nearest’、 ‘linear’、‘bilinear’、‘bicubic’、‘trilinear’，默认是’nearest’。上采样算法在本文中会有详细理论进行讲解；
④align_corners：如果True，输入和输出张量的角像素对齐，从而保留这些像素的值，默认是False。此处True和False的区别本文中会有详细的理论讲解；
⑤recompute_scale_factor：如果recompute_scale_factor是True，则必须传入scale_factor并且scale_factor用于计算输出大小。计算出的输出大小将用于推断插值的新比例。请注意，当scale_factor为浮点数时，由于舍入和精度问题，它可能与重新计算的scale_factor不同。如果recompute_scale_factor是False，那么size或scale_factor将直接用于插值。

class Up(nn.Module):
    def __init__(self):
        super().__init__()
        self.up = nn.Upsample(scale_factor=2, mode='bilinear', align_corners=True)

    def forward(self, x1, x2):
        x1 = self.up(x1)
        # input is CHW
        diffY = torch.tensor([x2.size()[2] - x1.size()[2]])
        diffX = torch.tensor([x2.size()[3] - x1.size()[3]])

        x1 = F.pad(x1, [diffX // 2, diffX - diffX // 2,
                        diffY // 2, diffY - diffY // 2])
        x = torch.cat([x2, x1], dim=1)
        return x
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

下采样

图中的黑色箭头，采用的是最大池化。

self.pool = nn.MaxPool2d(2, 2)
1

深度监督

所示，该结构下有4个分支，可以分为两种模式。

精确模式：4个分支取平均值结果

快速模式：只选择一个分支，其余被剪枝

if self.deep_supervision:
   output1 = self.final1(x0_1)
   output2 = self.final2(x0_2)
   output3 = self.final3(x0_3)
   output4 = self.final4(x0_4)
   return [output1, output2, output3, output4]

else:
    output = self.final(x0_4)
     return output
1
2
3
4
5
6
7
8
9
10

网络架构代码

class NestedUNet(nn.Module):
    def __init__(self, num_classes=1, input_channels=1, deep_supervision=False, **kwargs):
        super().__init__()

        nb_filter = [32, 64, 128, 256, 512]

        self.deep_supervision = deep_supervision

        self.pool = nn.MaxPool2d(2, 2)
        self.up = Up()
  
        self.conv0_0 = VGGBlock(input_channels, nb_filter[0], nb_filter[0])
        self.conv1_0 = VGGBlock(nb_filter[0], nb_filter[1], nb_filter[1])
        self.conv2_0 = VGGBlock(nb_filter[1], nb_filter[2], nb_filter[2])
        self.conv3_0 = VGGBlock(nb_filter[2], nb_filter[3], nb_filter[3])
        self.conv4_0 = VGGBlock(nb_filter[3], nb_filter[4], nb_filter[4])

        self.conv0_1 = VGGBlock(nb_filter[0]+nb_filter[1], nb_filter[0], nb_filter[0])
        self.conv1_1 = VGGBlock(nb_filter[1]+nb_filter[2], nb_filter[1], nb_filter[1])
        self.conv2_1 = VGGBlock(nb_filter[2]+nb_filter[3], nb_filter[2], nb_filter[2])
        self.conv3_1 = VGGBlock(nb_filter[3]+nb_filter[4], nb_filter[3], nb_filter[3])

        self.conv0_2 = VGGBlock(nb_filter[0]*2+nb_filter[1], nb_filter[0], nb_filter[0])
        self.conv1_2 = VGGBlock(nb_filter[1]*2+nb_filter[2], nb_filter[1], nb_filter[1])
        self.conv2_2 = VGGBlock(nb_filter[2]*2+nb_filter[3], nb_filter[2], nb_filter[2])

        self.conv0_3 = VGGBlock(nb_filter[0]*3+nb_filter[1], nb_filter[0], nb_filter[0])
        self.conv1_3 = VGGBlock(nb_filter[1]*3+nb_filter[2], nb_filter[1], nb_filter[1])

        self.conv0_4 = VGGBlock(nb_filter[0]*4+nb_filter[1], nb_filter[0], nb_filter[0])

        if self.deep_supervision:
            self.final1 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
            self.final2 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
            self.final3 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
            self.final4 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
        else:
            self.final = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)


    def forward(self, input):
        x0_0 = self.conv0_0(input)
        x1_0 = self.conv1_0(self.pool(x0_0))
        x0_1 = self.conv0_1(self.up(x1_0, x0_0))

        x2_0 = self.conv2_0(self.pool(x1_0))
        x1_1 = self.conv1_1(self.up(x2_0, x1_0))
        x0_2 = self.conv0_2(self.up(x1_1, torch.cat([x0_0, x0_1], 1)))

        x3_0 = self.conv3_0(self.pool(x2_0))
        x2_1 = self.conv2_1(self.up(x3_0, x2_0))   
        x1_2 = self.conv1_2(self.up(x2_1, torch.cat([x1_0, x1_1], 1)))
        x0_3 = self.conv0_3(self.up(x1_2, torch.cat([x0_0, x0_1, x0_2], 1)))

        x4_0 = self.conv4_0(self.pool(x3_0))
        x3_1 = self.conv3_1(self.up(x4_0, x3_0))
        x2_2 = self.conv2_2(self.up(x3_1, torch.cat([x2_0, x2_1], 1)))
        x1_3 = self.conv1_3(self.up(x2_2, torch.cat([x1_0, x1_1, x1_2], 1)))
        x0_4 = self.conv0_4(self.up(x1_3, torch.cat([x0_0, x0_1, x0_2, x0_3], 1)))

        if self.deep_supervision:
            output1 = self.final1(x0_1)
            output2 = self.final2(x0_2)
            output3 = self.final3(x0_3)
            output4 = self.final4(x0_4)
            return [output1, output2, output3, output4]

        else:
            output = self.final(x0_4)
            return output


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