为什么unet有类?

Why does unet have classes?

import torch
import torch.nn as nn
import torch.nn.functional as F


class double_conv(nn.Module):
    '''(conv => BN => ReLU) * 2'''
    def __init__(self, in_ch, out_ch):
        super(double_conv, self).__init__()
        self.conv = nn.Sequential(
            nn.Conv2d(in_ch, out_ch, 3, padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU(inplace=True),
            nn.Conv2d(out_ch, out_ch, 3, padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU(inplace=True)
        )

    def forward(self, x):
        x = self.conv(x)
        return x


class inconv(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(inconv, self).__init__()
        self.conv = double_conv(in_ch, out_ch)

    def forward(self, x):
        x = self.conv(x)
        return x


class down(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(down, self).__init__()
        self.mpconv = nn.Sequential(
            nn.MaxPool2d(2),
            double_conv(in_ch, out_ch)
        )

    def forward(self, x):
        x = self.mpconv(x)
        return x


class up(nn.Module):
    def __init__(self, in_ch, out_ch, bilinear=True):
        super(up, self).__init__()

        #  would be a nice idea if the upsampling could be learned too,
        #  but my machine do not have enough memory to handle all those weights
        if bilinear:
            self.up = nn.Upsample(scale_factor=2, mode='bilinear', align_corners=True)
        else:
            self.up = nn.ConvTranspose2d(in_ch//2, in_ch//2, 2, stride=2)

        self.conv = double_conv(in_ch, out_ch)

    def forward(self, x1, x2):
        x1 = self.up(x1)
        diffX = x1.size()[2] - x2.size()[2]
        diffY = x1.size()[3] - x2.size()[3]
        x2 = F.pad(x2, (diffX // 2, int(diffX / 2),
                        diffY // 2, int(diffY / 2)))
        x = torch.cat([x2, x1], dim=1)
        x = self.conv(x)
        return x


class outconv(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(outconv, self).__init__()
        self.conv = nn.Conv2d(in_ch, out_ch, 1)

    def forward(self, x):
        x = self.conv(x)
        return x


class UNet(nn.Module):
    def __init__(self, n_channels, n_classes):
        super(UNet, self).__init__()
        self.inc = inconv(n_channels, 64)
        self.down1 = down(64, 128)
        self.down2 = down(128, 256)
        self.down3 = down(256, 512)
        self.down4 = down(512, 512)
        self.up1 = up(1024, 256)
        self.up2 = up(512, 128)
        self.up3 = up(256, 64)
        self.up4 = up(128, 64)
        self.outc = outconv(64, n_classes)

    def forward(self, x):
        self.x1 = self.inc(x)
        self.x2 = self.down1(self.x1)
        self.x3 = self.down2(self.x2)
        self.x4 = self.down3(self.x3)
        self.x5 = self.down4(self.x4)
        self.x6 = self.up1(self.x5, self.x4)
        self.x7 = self.up2(self.x6, self.x3)
        self.x8 = self.up3(self.x7, self.x2)
        self.x9 = self.up4(self.x8, self.x1)
        self.y = self.outc(self.x9)
        return self.y

当我阅读 UNet 架构时,我发现它有 n_classes 作为输出。

class UNet(nn.Module):
    def __init__(self, n_channels, n_classes):

但是为什么它有n_classes因为它是用于图像分割的?

我正在尝试使用此代码进行图像去噪,但我无法弄清楚 n_classes 参数应该是什么,因为我没有任何 类.

n_classes是否表示多类分割?如果是这样,二进制UNet分割的输出是什么?

回答

Does n_classes signify multiclass segmentation?

是的,如果你指定n_classes=4它会输出一个(batch, 4, width, height)形状的张量,其中每个像素都可以被分割为4类之一。还应该使用 torch.nn.CrossEntropyLoss 进行训练。

If so, what is the output of binary UNet segmentation?

如果你想使用二进制分割,你会指定 n_classes=10 表示黑色或 1 表示白色)并使用 torch.nn.BCEWithLogitsLoss

I am trying to use this code for image denoising and I couldn't figure out what will should the n_classes parameter be

应该等于n_channels,通常RGB是3,灰度是1。如果你想教这个模型去噪你应该:

  • 给图像添加一些噪点(例如使用 torchvision.transforms
  • 在最后使用 sigmoid 激活,因为像素的值介于 01 之间(除非标准化)
  • 使用torch.nn.MSELoss进行训练

为什么是 sigmoid?

因为 [0,255] 像素范围表示为 [0, 1] 像素值(至少没有归一化)。 sigmoid 确实如此 - 将值压缩到 [0, 1] 范围内,因此 linear 输出(logits)的范围可以从 -inf+inf.

Why not a linear output and a clamp?

为了使线性层在钳位后处于 [0, 1] 范围内,线性层的可能输出值必须大于 0(适合目标的对数范围:[0, +inf])

Why not a linear output without a clamp?

输出的 Logits 必须在 [0, 1] 范围内

Why not some other method?

你可以这样做,但是 sigmoid 的想法是:

  • 帮助神经网络(可以输出任意logit值
  • sigmoid 的一阶导数是高斯标准正态分布,因此它模拟了许多现实生活中发生的现象的概率(更多信息请参见 here