RBF多尺度融合卷积，保持尺寸不变，融合各个维度Pytorch

本文链接：https://blog.csdn.net/qq_26974921/article/details/130769492

该代码定义了一个基础的卷积模块BasicConv2d，以及一个RFB_modified模块，该模块包含了四个不同膨胀率的卷积分支，用于多尺度特征提取，最后通过拼接和通道压缩得到输出。在主函数中，创建了RFB_modified实例并进行了测试运算。

摘要由CSDN通过智能技术生成

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



class BasicConv2d(nn.Module):
    """
    这是一个基础的卷积模块，可进行参数设置，膨胀卷积和其他参数
    """
    def __init__(self, in_planes, out_planes, kernel_size, stride=1, padding=0, dilation=1):
        super(BasicConv2d, self).__init__()
        self.conv = nn.Conv2d(in_planes, out_planes,
                              kernel_size=kernel_size, stride=stride,
                              padding=padding, dilation=dilation, bias=False)
        self.bn = nn.BatchNorm2d(out_planes)
        self.relu = nn.ReLU(inplace=True)

    def forward(self, x):
        x = self.conv(x)
        x = self.bn(x)
        return x
class RFB_modified(nn.Module):
    def __init__(self, in_channel, out_channel):
        super(RFB_modified, self).__init__()
        self.relu = nn.ReLU(True)
        self.branch0 = nn.Sequential(
            BasicConv2d(in_channel, out_channel, 1),
        )#通道变换
        self.branch1 = nn.Sequential(
            BasicConv2d(in_channel, out_channel, 1),
            BasicConv2d(out_channel, out_channel, kernel_size=(1, 3), padding=(0, 1)),
            BasicConv2d(out_channel, out_channel, kernel_size=(3, 1), padding=(1, 0)),
            BasicConv2d(out_channel, out_channel, 3, padding=3, dilation=3)
        )#第一分支不变尺寸卷积
        self.branch2 = nn.Sequential(
            BasicConv2d(in_channel, out_channel, 1),
            BasicConv2d(out_channel, out_channel, kernel_size=(1, 5), padding=(0, 2)),
            BasicConv2d(out_channel, out_channel, kernel_size=(5, 1), padding=(2, 0)),
            BasicConv2d(out_channel, out_channel, 3, padding=5, dilation=5)
        )##第二分支不变尺寸卷积
        self.branch3 = nn.Sequential(
            BasicConv2d(in_channel, out_channel, 1),
            BasicConv2d(out_channel, out_channel, kernel_size=(1, 7), padding=(0, 3)),
            BasicConv2d(out_channel, out_channel, kernel_size=(7, 1), padding=(3, 0)),
            BasicConv2d(out_channel, out_channel, 3, padding=7, dilation=7)
        )#第三分支不变尺寸卷积
        self.conv_cat = BasicConv2d(4*out_channel, out_channel, 3, padding=1)
        #多尺度拼接以后进行将通道
        self.conv_res = BasicConv2d(in_channel, out_channel, 1)
        #通道压缩
    def forward(self, x):
        x0 = self.branch0(x)
        x1 = self.branch1(x)
        print(x1.shape)
        x2 = self.branch2(x)
        print(x2.shape)
        x3 = self.branch3(x)
        print(x3.shape)
        x_cat = self.conv_cat(torch.cat((x0, x1, x2, x3), 1))#多尺度拼接
        print(x_cat.shape)
        x = self.relu(x_cat + self.conv_res(x))
        return x
if __name__ == '__main__':
    ras = RFB_modified(1,1).cuda()
    input_tensor = torch.randn(1, 1, 352, 352).cuda()
    out = ras(input_tensor)
    print(out)
    print(out.shape)