YOLOV5改进（DCN）

DCN（可变形卷积）

1.适用于长宽比非常悬殊,且一般加到主干网络（Backbone）上

2.DCN讲解：https://blog.csdn.net/weixin_44227733/article/details/124392777?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522168610235316800225572865%2522%252C%2522scm%2522%253A%252220140713.130102334..%2522%257D&request_id=168610235316800225572865&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2~all~top_positive~default-1-124392777-null-null.142^v88^koosearch_v1,239^v2^insert_chatgpt&utm_term=PPYOLO&spm=1018.2226.3001.4187&ydreferer=aHR0cHM6Ly9zby5jc2RuLm5ldC9zby9zZWFyY2g%2FcT1QUFlPTE8mdD0mdT0mdXJ3PQ%3D%3D

（1）PP-YOLO使用DCN的方法

可以看出只是将最后一层卷积换成DCN

（2）Cascade-RCNN中对DCN的使用

 （False,True,True,True）可以看出其对DCN的使用和PP-YOLO类似就是后面三个卷积替换成DCN

 （3）为什么要使用可变形卷积

 

3.适用方法：

（1）将下方代码复制到models/commen.py中

class DCNv2(nn.Module):
    def __init__(self, in_channels, out_channels, kernel_size, stride=1,
                 padding=1, dilation=1, groups=1, deformable_groups=1):
        super(DCNv2, self).__init__()

        self.in_channels = in_channels
        self.out_channels = out_channels
        self.kernel_size = (kernel_size, kernel_size)
        self.stride = (stride, stride)
        self.padding = (padding, padding)
        self.dilation = (dilation, dilation)
        self.groups = groups
        self.deformable_groups = deformable_groups

        self.weight = nn.Parameter(
            torch.empty(out_channels, in_channels, *self.kernel_size)
        )
        self.bias = nn.Parameter(torch.empty(out_channels))

        out_channels_offset_mask = (self.deformable_groups * 3 *
                                    self.kernel_size[0] * self.kernel_size[1])
        self.conv_offset_mask = nn.Conv2d(
            self.in_channels,
            out_channels_offset_mask,
            kernel_size=self.kernel_size,
            stride=self.stride,
            padding=self.padding,
            bias=True,
        )
        self.bn = nn.BatchNorm2d(out_channels)
        self.act = Conv.default_act
        self.reset_parameters()

    def forward(self, x):
        offset_mask = self.conv_offset_mask(x)
        o1, o2, mask = torch.chunk(offset_mask, 3, dim=1)
        offset = torch.cat((o1, o2), dim=1)
        mask = torch.sigmoid(mask)
        x = torch.ops.torchvision.deform_conv2d(
            x,
            self.weight,
            offset,
            mask,
            self.bias,
            self.stride[0], self.stride[1],
            self.padding[0], self.padding[1],
            self.dilation[0], self.dilation[1],
            self.groups,
            self.deformable_groups,
            True
        )
        x = self.bn(x)
        x = self.act(x)
        return x

    def reset_parameters(self):
        n = self.in_channels
        for k in self.kernel_size:
            n *= k
        std = 1. / math.sqrt(n)
        self.weight.data.uniform_(-std, std)
        self.bias.data.zero_()
        self.conv_offset_mask.weight.data.zero_()
        self.conv_offset_mask.bias.data.zero_()

class Bottleneck_DCN(nn.Module):
    # Standard bottleneck
    def __init__(self, c1, c2, shortcut=True, g=1, e=0.5):  # ch_in, ch_out, shortcut, groups, expansion
        super().__init__()
        c_ = int(c2 * e)  # hidden channels
        self.cv1 = Conv(c1, c_, 1, 1)
        self.cv2 = DCNv2(c_, c2, 3, 1, groups=g)
        self.add = shortcut and c1 == c2

    def forward(self, x):
        return x + self.cv2(self.cv1(x)) if self.add else self.cv2(self.cv1(x))

class C3_DCN(C3):
    # C3 module with DCNv2
    def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5):
        super().__init__(c1, c2, n, shortcut, g, e)
        c_ = int(c2 * e)
        self.m = nn.Sequential(*(Bottleneck_DCN(c_, c_, shortcut, g, e=1.0) for _ in range(n)))

（2）在models/yolo.py中的解析模型中增加C3_DCN

 （3）更换.yaml中backbone的C3。（四个block替换后三个block）

 （4）运行