英文:
How to apply Fully Convolutional Network (FCN) to binary classification?
问题
To change out_put_1.shape to [1, 1], you can modify the FCNDenseNet121_1 model's final classifier layer. Currently, it has a nn.Conv3d layer with a kernel size of 1, which results in the shape [1, 1, 3, 3, 3]. To change it to [1, 1], you should update the kernel size and stride of the nn.Conv3d layer to 3, like this:
self.classifier = nn.Conv3d(out_channels, 1, kernel_size=3, stride=3)
This will produce an output shape of [1, 1].
Regarding whether this is the correct way, it depends on your specific problem and requirements. Using a 3x3x3 convolution with a stride of 3 effectively reduces the spatial dimensions and channel count, resulting in a single-channel output. Whether this is suitable for your binary classification task depends on the nature of your data and the problem you're trying to solve.
As for out_put_2.shape being [1, 1], if it meets your requirements and provides the desired results for your binary classification task, it can be considered a valid approach. However, you should evaluate both models' performance on your dataset to determine which one works better for your specific problem.
英文:
I want to do binary classification with FCN. This is the code I implemented.
import monai
import torch
import torch.nn as nn
#FCN + DenseNet
class FCNDenseNet121_1(nn.Module):
def __init__(self, spatial_dims:int, in_channels:int, out_channels:int):
super().__init__()
self.encoder = monai.networks.nets.DenseNet121(spatial_dims=spatial_dims, in_channels=in_channels, out_channels=out_channels)
self.encoder.class_layers.pool = nn.AvgPool3d(kernel_size=1, stride=1)
self.encoder.class_layers.flatten = nn.Identity()
self.encoder.class_layers.out = nn.Identity()
self.classifier = nn.Conv3d(out_channels, 1, kernel_size=1, stride=1)
def forward(self, x):
x = self.encoder(x)
x = self.classifier(x)
return x
class FCNDenseNet121_2(nn.Module):
def __init__(self, spatial_dims:int, in_channels:int, out_channels:int):
super().__init__()
self.encoder = monai.networks.nets.DenseNet121(spatial_dims=spatial_dims, in_channels=in_channels, out_channels=out_channels)
#self.encoder.class_layers.pool = nn.AvgPool3d(kernel_size=1, stride=1)
self.encoder.class_layers.flatten = nn.Identity()
self.encoder.class_layers.out = nn.Identity()
self.classifier = nn.Conv3d(out_channels, 1, kernel_size=1, stride=1)
def forward(self, x):
x = self.encoder(x)
x = self.classifier(x)
x = x.view(x.size(0), -1)
return x
#Built a model, my images are 3 dimensions
model_1 = FCNDenseNet121_1(spatial_dims=3, in_channels=1, out_channels=1024)
model_2 = FCNDenseNet121_2(spatial_dims=3, in_channels=1, out_channels=1024)
loss_function = torch.nn.BCEWithLogitsLoss()
optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
#The operation verification
x = torch.randn(1,96,96,96)
print(x.shape) #torch.Size([1, 96, 96, 96])
out_put_1 = model_1(x.unsqueeze(0))
out_put_2 = model_2(x.unsqueeze(0))
print(out_put_1.shape) #torch.Size([1, 1, 3, 3, 3])
print(out_put_2.shape) #torch.Size([1, 1])
I want to ask how to change out_put_1.shape to [1,1] to calculate loss.
First of all, is this correct way?
Although out_put_2.shape is [1,1], I don't know if this way of doing it is good.
Which is correct?
答案1
得分: 1
如果您的目标是执行二元分类，并且希望输出具有形状[1, 1]，则应该使用 Model 2（FCNDenseNet121_2）。这是因为它已经输出了一个具有所需形状的张量。
但是，为了计算损失，您需要确保目标张量也具有形状[1, 1]，以匹配输出张量。例如，如果您的目标张量是y，它也应该具有形状[1, 1]：
y = torch.tensor([[1]]) # 具有形状[1, 1]的示例目标张量
loss = loss_function(out_put_2, y)
请记住，在训练模型时，您还应该修改批量大小，因为您当前使用的批量大小为1。这通常不建议，因为它可能导致训练不稳定。增加批量大小以获得更好的训练性能。
英文:
If your goal is to perform binary classification, and you want your output to have the shape [1, 1], you should use Model 2 (FCNDenseNet121_2). This is because it already outputs a tensor with the desired shape.
However, to calculate the loss, you need to make sure the target tensor also has a shape of [1, 1] to match the output tensor. For instance, if your target tensor is y, it should have a shape of [1, 1] as well:
y = torch.tensor([[1]]) # example target tensor with shape [1, 1]
loss = loss_function(out_put_2, y)
Keep in mind that you should also modify the batch size when training the model, as you're currently using a batch size of 1. This is generally not recommended, as it can lead to unstable training. Increase the batch size for better training performance.
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论