DCA特征融合的pytorch实现

以下是DCA特征融合的pytorch实现：

import torch
import torch.nn as nn
import torch.nn.functional as F

class DCA(nn.Module):
    def __init__(self, in_channels):
        super(DCA, self).__init__()
        self.conv1 = nn.Conv2d(in_channels=in_channels, out_channels=in_channels, kernel_size=3, stride=1, padding=1)
        self.conv2 = nn.Conv2d(in_channels=in_channels, out_channels=in_channels, kernel_size=3, stride=1, padding=1)
        self.conv3 = nn.Conv2d(in_channels=in_channels, out_channels=in_channels, kernel_size=3, stride=1, padding=1)
        self.conv4 = nn.Conv2d(in_channels=in_channels, out_channels=in_channels, kernel_size=3, stride=1, padding=1)

    def forward(self, x):
        x1 = self.conv1(x)
        x1 = F.relu(x1)
        x1 = self.conv2(x1)
        x1 = F.relu(x1)

        x2 = self.conv3(x)
        x2 = F.relu(x2)
        x2 = self.conv4(x2)
        x2 = F.relu(x2)

        x = x + x1 + x2
        return x

在这里，我们定义了一个DCA类，该类继承了nn.Module。在类的构造函数中，我们定义了四个卷积层，用于提取特征。在前向方法中，我们首先将输入张量传递到第一个卷积层，然后使用ReLU激活函数进行激活，再将结果传递到第二个卷积层，再次使用ReLU激活函数进行激活。同样，我们使用第三个和第四个卷积层提取另一个不同的特征。最后，我们将这三个特征加在一起，以产生最终特征表示。