三分钟学会使用系列（YOLOv5）|SK通道注意力机制，涨点神器！

三分钟学会使用系列（YOLOv5）|SK通道注意力机制，涨点神器！

---

本文旨在让读者快速了解相关技术并运用,欢迎关注。

1. 原文地址：《Selective Kernel Networks》

---

1. SK通道注意力导读

​ Selective Kernel（SK）使用多个不同分支中的引导信息进行softmax attention融合。对这些分支的卷积核大小不同，因此融合层中神经元的有效感受野的大小不同。通俗的讲，Sk的作者提出了一种动态选择机制，允许每个神经元根据输入信息的多个尺度自适应地调整其感受野大小。

---

2. 代码

​ SK代码。

class SKConv(nn.Module):
    def __init__(self, in_channel, WH=1, M=2, G=1, r=2, stride=1, L=32):
        super(SKConv, self).__init__()
        d = max(int(in_channel / r), L)
        self.M = M
        self.features = features
        self.convs = nn.ModuleList([])
        for i in range(M):
            self.convs.append(
                nn.Sequential(
                    nn.Conv2d(in_channel,
                              in_channel,
                              kernel_size=3 + i * 2,
                              stride=stride,
                              padding=1 + i,
                              groups=G), nn.BatchNorm2d(in_channel),
                    nn.ReLU(inplace=False)))
            
        self.fc = nn.Linear(in_channel, d)
        self.fcs = nn.ModuleList([])
        for i in range(M):
            self.fcs.append(nn.Linear(d, in_channel))
        self.softmax = nn.Softmax(dim=1)

    def forward(self, x):
        for i, conv in enumerate(self.convs):
            fea = conv(x).unsqueeze_(dim=1)
            if i == 0:
                feas = fea
            else:
                feas = torch.cat([feas, fea], dim=1)
        fea_U = torch.sum(feas, dim=1)
        fea_s = fea_U.mean(-1).mean(-1)
        fea_z = self.fc(fea_s)
        for i, fc in enumerate(self.fcs):
            # print(i, fea_z.shape)
            vector = fc(fea_z).unsqueeze_(dim=1)
            # print(i, vector.shape)
            if i == 0:
                attention_vectors = vector
            else:
                attention_vectors = torch.cat([attention_vectors, vector],
                                              dim=1)
        attention_vectors = self.softmax(attention_vectors)
        attention_vectors = attention_vectors.unsqueeze(-1).unsqueeze(-1)
        fea_v = (feas * attention_vectors).sum(dim=1)
        return fea_v

---

3. 使用教程

​ 以YOLOv5为例，加入SK注意力机制。

# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license

# Parameters
nc: 80  # number of classes
depth_multiple: 0.33  # model depth multiple
width_multiple: 0.25  # layer channel multiple
anchors:
  - [10,13, 16,30, 33,23]  # P3/8
  - [30,61, 62,45, 59,119]  # P4/16
  - [116,90, 156,198, 373,326]  # P5/32

# YOLOv5 v6.0 backbone
backbone:
  # [from, number, module, args]
  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
   [-1, 3, C3, [128]],
   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
   [-1, 6, C3, [256]],
   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
   [-1, 9, C3, [512]],
   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
   [-1, 3, C3, [1024]],
   [-1, 1, SPPF, [1024, 5]],  # 9
  ]

# YOLOv5 v6.0 head
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 13

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)

   [-1, 1, Conv, [256, 3, 2]],
   [[-1, 14], 1, Concat, [1]],  # cat head P4
   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)
   [-1, 1, SKConv, []],  # 24

   [[17, 20, 24], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]

---

4. 参考

1. 原文地址：《Selective Kernel Networks》

---