CNN中的注意力机制

SE

Squeeze-and-Excitation Networks。

通过全局平均池化，先降维再升维的变换，使用sigmoid函数实现特征的重标定，给通道赋值不同的权重，最后再将权重加载到输入中。

class SELayer(nn.Module):
    def __init__(self, channel, reduction=16):
        super(SELayer, self).__init__()
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.fc = nn.Sequential(
            nn.Linear(channel, channel // reduction, bias=False),
            nn.ReLU(inplace=True),
            nn.Linear(channel // reduction, channel, bias=False),
            nn.Sigmoid()
        )

    def forward(self, x):
        b, c, _, _ = x.size()
        y = self.avg_pool(x).view(b, c)
        y = self.fc(y).view(b, c, 1, 1)
        return x * y.expand_as(x)

SE模块和残差模块结合，一般放在残差分支，和输入相加之前，可以对残差分支学习到的特征进行加权调整，从而增强高级特征的表达能力。

class SEBasicBlock(nn.Module):
    expansion = 1

    def __init__(self, inplanes, planes, stride=1, downsample=None, groups=1,
                 base_width=64, dilation=1, norm_layer=None,
                 *, reduction=16):
        super(SEBasicBlock, self).__init__()
        self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=3,stride=stride, padding=1)
        self.bn1 = nn.BatchNorm2d(planes)
        self.relu = nn.ReLU(inplace=True)
        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3,stride=1, padding=1)
        self.bn2 = nn.BatchNorm2d(planes)
        self.se = SELayer(planes, reduction)
        self.downsample = downsample
        self.stride = stride

    def forward(self, x):
        residual = x
        out = self.relu(self.bn1(self.conv1(x)))
        out = self.se(self.bn2(self.conv2(out)))  #use SE

        if self.downsample is not None:
            residual = self.downsample(x)

        out += residual
        out = self.relu(out)

        return out

 ECA

Effificient Channel Attention for Deep Convolutional Neural Networks。

作者经过实验研究表明，SENet中降维对通道的注意力预测带来了副作用，提出了一种不降维的局部跨通道交互策略，使用一维卷积替代了SENet中的MLP。

class ECALayer(nn.Module):
    """Constructs a ECA module.
    Args:
        channel: Number of channels of the input feature map
        k_size: Adaptive selection of kernel size
    """
    def __init__(self, channel, k_size=3):
        super(eca_layer, self).__init__()
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.conv = nn.Conv1d(1, 1, kernel_size=k_size, padding=(k_size - 1) // 2, bias=False)
        self.sigmoid = nn.Sigmoid()
 
    def forward(self, x):
        # x: input features with shape [b, c, h, w]
        b, c, h, w = x.size()
 
        # feature descriptor on the global spatial information
        y = self.avg_pool(x)   #  [b, c, 1, 1]
 
        # Two different branches of ECA module
        y = y.squeeze(-1).transpose(-1, -2) # [b, c, 1] => [b, 1,c]
        y = self.conv(y)  # [b, 1,c]
        y = y.transpose(-1, -2).unsqueeze(-1) # [b, c, 1] => [b, c, 1, 1]
 
        # Multi-scale information fusion
        y = self.sigmoid(y)
        
        return x * y.expand_as(x)

CA 

Coordinate attention for efficient mobile