深度学习-注意力机制

最新推荐文章于 2024-04-28 19:44:21 发布

DotNet的小云

最新推荐文章于 2024-04-28 19:44:21 发布

阅读量965

点赞数 21

文章标签：深度学习

本文链接：https://blog.csdn.net/qq_45085522/article/details/134559507

版权

基于MK-MMD度量迁移学习的轴承故障诊断方法研究-CSDN博客

注意力机制

一、注意力模块

神经网络学习小记录64——Pytorch 图像处理中注意力机制的解析与代码详解_https://blog.csdn.net/weixin_44791964/article/deta_Bubbliiiing的博客-CSDN博客

1.1、SEnet模块

nn.AdaptiveAvgPool2d(1)

Pytorch学习笔记(五):nn.AdaptiveAvgPool2d()函数详解_nn.adaptiveavgpool2d((1, 1))-CSDN博客

含义：

注意力机制在CNN中使用总结_注意力机制加在cnn的什么位置_AI浩的博客-CSDN博客

import torch
import torch.nn as nn
m = nn.AdaptiveAvgPool2d((5,1))
m1 = nn.AdaptiveAvgPool2d((None,5))
m2 = nn.AdaptiveAvgPool2d(1)
input = torch.randn(2, 64, 8, 9)
output = m(input)
output1 = m1(input)
output2 = m2(input)
print('nn.AdaptiveAvgPool2d((5,1)):',output.shape)
print('nn.AdaptiveAvgPool2d((None,5)):',output1.shape)
print('nn.AdaptiveAvgPool2d(1):',output2.shape)


class senet(nn. Module):
    def _init__(self,channel,ratio = 16):
        super(senet,self)._init__()
        self.avg.pool= nn. AdaptiveAvgPool2d(1)
        self.fc= nn.Sequential(
        nn.Linear(channel,channel//ratio,False),
        nn.ReLU(),
        nn.Linear(channel//ratio,channel,False),
        nn.Sigmoid(),
        )
    def forward(self,x):
        b, c, h, w = x.size()
        # b, c, h, w ->b, c,1,1
        avg = self.avg_pool(x).view([b,c])
        # b,c ->b,c l/ ratio ->b, c ->b,c，1，1
        fc = self.fc(avg).view([b,c,1,1])
        return x*fc



model = senet(512)
print(model)
inputs = torch.ones ([2,512,26,26])
outputs = model(inputs)
print(outputs.shape)

1.2、CBAM模块实现

CBAM 通道注意力 & 空间注意力机制_空间注意力机制和通道注意力机制-CSDN博客

通道和空间注意力机制的结合

import torch  
import torch.nn as nn  
import torch.nn.functional as F  
  
class CBAM(nn.Module):  
    def __init__(self, channels, reduction):  
        super(CBAM, self).__init__()  
        self.avg_pool = nn.AdaptiveAvgPool2d(1)  
        self.max_pool = nn.AdaptiveMaxPool2d(1)  
        self.fc = nn.Sequential(  
            nn.Linear(channels, channels // reduction, bias=False),  
            nn.ReLU(inplace=True),  
            nn.Linear(channels // reduction, channels, bias=False),  
            nn.Sigmoid()  
        )  
    def forward(self, x):  
        b, c, _, _ = x.size()  
        y = self.avg_pool(x).view(b, c)  
        z = self.max_pool(x).view(b, c)  
        y = self.fc(y)  
        z = self.fc(z)  
        x = x * y * z.expand_as(x)  
        return x

应用：如何应用注意力机制

class MyModel(nn.Module):  
    def __init__(self):  
        super(MyModel, self).__init__()  
        self.conv1 = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1)  
        self.cbam = CBAM(64, 16)  # CBAM模块的输入通道数是64，reduction参数是16  
        self.conv2 = nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1)  
    def forward(self, x):  
        x = F.relu(self.conv1(x))  
        x = self.cbam(x)  # 在卷积后插入CBAM模块  
        x = F.relu(self.conv2(x))  
        return x

class ChannelAttention(nn.Module):
    def __init__(self, in_planes, ratio=8):
        super(ChannelAttention, self).__init__()
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.max_pool = nn.AdaptiveMaxPool2d(1)

        # 利用1x1卷积代替全连接
        self.fc1   = nn.Conv2d(in_planes, in_planes // ratio, 1, bias=False)
        self.relu1 = nn.ReLU()
        self.fc2   = nn.Conv2d(in_planes // ratio, in_planes, 1, bias=False)

        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        avg_out = self.fc2(self.relu1(self.fc1(self.avg_pool(x))))
        max_out = self.fc2(self.relu1(self.fc1(self.max_pool(x))))
        out = avg_out + max_out
        return self.sigmoid(out)

class SpatialAttention(nn.Module):
    def __init__(self, kernel_size=7):
        super(SpatialAttention, self).__init__()

        assert kernel_size in (3, 7), 'kernel size must be 3 or 7'
        padding = 3 if kernel_size == 7 else 1
        self.conv1 = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=False)
        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        avg_out = torch.mean(x, dim=1, keepdim=True)
        max_out, _ = torch.max(x, dim=1, keepdim=True)
        x = torch.cat([avg_out, max_out], dim=1)
        x = self.conv1(x)
        return self.sigmoid(x)

class cbam_block(nn.Module):
    def __init__(self, channel, ratio=8, kernel_size=7):
        super(cbam_block, self).__init__()
        self.channelattention = ChannelAttention(channel, ratio=ratio)
        self.spatialattention = SpatialAttention(kernel_size=kernel_size)

    def forward(self, x):
        x = x * self.channelattention(x)
        x = x * self.spatialattention(x)
        return x

由于在大部分情况下我们感兴趣的区域只是图像中的一小部分，因此空间注意力的本质就是定位目标并进行一些变换或者获取权重

1.3、ECAnet注意力机制实现