【目标检测】|ASFF

class ASFF(nn.Module):def __init__(self, level, rfb=False, vis=False):super(ASFF, self).__init__()self.level = level# 输入的三个特征层的channels, 根据实际修改self.dim = [512, 256, 256]self.inter_dim = self.dim[self.

rrr2

559人浏览 · 2021-06-18 22:53:10

rrr2 · 2021-06-18 22:53:10 发布

在这里插入图片描述

class ASFF(nn.Module):
    def __init__(self, level, rfb=False, vis=False):
        super(ASFF, self).__init__()
        self.level = level
        # 输入的三个特征层的channels, 根据实际修改
        self.dim = [512, 256, 256]
        self.inter_dim = self.dim[self.level]
        # 每个层级三者输出通道数需要一致
        if level==0:
            self.stride_level_1 = add_conv(self.dim[1], self.inter_dim, 3, 2)
            self.stride_level_2 = add_conv(self.dim[2], self.inter_dim, 3, 2)
            self.expand = add_conv(self.inter_dim, 1024, 3, 1)
        elif level==1:
            self.compress_level_0 = add_conv(self.dim[0], self.inter_dim, 1, 1)
            self.stride_level_2 = add_conv(self.dim[2], self.inter_dim, 3, 2)
            self.expand = add_conv(self.inter_dim, 512, 3, 1)
        elif level==2:
            self.compress_level_0 = add_conv(self.dim[0], self.inter_dim, 1, 1)
            if self.dim[1] != self.dim[2]:
                self.compress_level_1 = add_conv(self.dim[1], self.inter_dim, 1, 1)
            self.expand = add_conv(self.inter_dim, 256, 3, 1)

        compress_c = 8 if rfb else 16  #when adding rfb, we use half number of channels to save memory

        self.weight_level_0 = add_conv(self.inter_dim, compress_c, 1, 1)
        self.weight_level_1 = add_conv(self.inter_dim, compress_c, 1, 1)
        self.weight_level_2 = add_conv(self.inter_dim, compress_c, 1, 1)

        self.weight_levels = nn.Conv2d(compress_c*3, 3, kernel_size=1, stride=1, padding=0)
        self.vis= vis

	# 尺度大小 level_0 < level_1 < level_2
    def forward(self, x_level_0, x_level_1, x_level_2):
        if self.level==0:
            level_0_resized = x_level_0
            level_1_resized = self.stride_level_1(x_level_1)

            level_2_downsampled_inter =F.max_pool2d(x_level_2, 3, stride=2, padding=1)
            level_2_resized = self.stride_level_2(level_2_downsampled_inter)

        elif self.level==1:
            level_0_compressed = self.compress_level_0(x_level_0)
            level_0_resized =F.interpolate(level_0_compressed, scale_factor=2, mode='nearest')
            level_1_resized =x_level_1
            level_2_resized =self.stride_level_2(x_level_2)
        elif self.level==2:
            level_0_compressed = self.compress_level_0(x_level_0)
            level_0_resized =F.interpolate(level_0_compressed, scale_factor=4, mode='nearest')
            if self.dim[1] != self.dim[2]:
                level_1_compressed = self.compress_level_1(x_level_1)
                level_1_resized = F.interpolate(level_1_compressed, scale_factor=2, mode='nearest')
            else:
                level_1_resized =F.interpolate(x_level_1, scale_factor=2, mode='nearest')
            level_2_resized =x_level_2

        level_0_weight_v = self.weight_level_0(level_0_resized)
        level_1_weight_v = self.weight_level_1(level_1_resized)
        level_2_weight_v = self.weight_level_2(level_2_resized)
        levels_weight_v = torch.cat((level_0_weight_v, level_1_weight_v, level_2_weight_v),1)
        levels_weight = self.weight_levels(levels_weight_v)
        levels_weight = F.softmax(levels_weight, dim=1)   # alpha等产生

        fused_out_reduced = level_0_resized * levels_weight[:,0:1,:,:]+\
                            level_1_resized * levels_weight[:,1:2,:,:]+\
                            level_2_resized * levels_weight[:,2:,:,:]

        out = self.expand(fused_out_reduced)

        if self.vis:
            return out, levels_weight, fused_out_reduced.sum(dim=1)
        else:
            return out