UNet改进(20):融合通道-空间稀疏注意力的医学图像分割模型
1. UNet架构回顾
UNet最初由Ronneberger等人提出,专门用于生物医学图像分割。其经典结构由对称的编码器(下采样)和解码器(上采样)路径组成,中间通过跳跃连接(skip connections)将低级特征与高级特征融合。
class UNet(nn.Module):def __init__(self, in_channels=1, num_classes=1):super(UNet, self).__init__()self.in_channels = in_channelsself.num_classes = num_classes# 编码器路径(下采样)self.in_conv = DoubleConv(in_channels, 64)self.down1 = Down(64, 128)self.down2 = Down(128, 256)self.down3 = Down(256, 512)self.down4 = Down(512, 1024)# 解码器路径(上采样)self.up1 = Up(1024, 512)self.up2 =