文章目录

00 写在前面

感知损失：对于提升图片的肉眼可见细节，效果十分明显；对于一些指标如（SSIM、PSNR）这些，效果不明显。

在01中，可以根据3D VGG的网络结构，进行模块化编程，主要包括VGG3D模块。

在模型调试过程中，可以先通过简单测试代码，进行代码调试。

01 基于Pytorch版本的3D VGG代码

# 库函数调用
import torch
import torch.nn as nn

# VGG3D模块
class CustomVGG3D(nn.Module):
    def __init__(self, in_channels=3, out_channels=2):
        super(CustomVGG3D, self).__init__()
        self.features = nn.Sequential(
            nn.Conv3d(in_channels, 64, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            nn.Conv3d(64, 64, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)),
            nn.Conv3d(64, 128, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            nn.Conv3d(128, 128, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)),
            nn.Conv3d(128, 256, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            nn.Conv3d(256, 256, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            nn.Conv3d(256, 256, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            nn.Conv3d(256, 256, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            nn.ReLU(inplace=True),
            # nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)),
            # nn.Conv3d(256, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            # nn.ReLU(inplace=True),
            # nn.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            # nn.ReLU(inplace=True),
            # nn.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            # nn.ReLU(inplace=True),
            # nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)),
            # nn.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            # nn.ReLU(inplace=True),
            # nn.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            # nn.ReLU(inplace=True),
            # nn.Conv3d(512, 512, kernel_size=(3, 3, 3), padding=(1, 1, 1)),
            # nn.ReLU(inplace=True),
            # nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2)),
        )
        self.classifier = nn.Sequential(
            nn.Linear(512 * 8 * 8 * 1, 4096),
            nn.ReLU(True),
            nn.Linear(4096, 4096),
            nn.ReLU(True),
            nn.Linear(4096, out_channels),
            nn.Sigmoid()
        )

    def forward(self, x):
        x = self.features(x)
        # x = x.view(x.size(0), -1)
        # x = self.classifier(x)
        return x

# 测试代码
# if __name__ == '__main__':
#     x = torch.ones([2, 4, 256, 256, 32])
#     model = CustomVGG3D(in_channels=4, out_channels=1)
#     f = model(x)
#     print(f)

02 论文下载

Very deep convolutional neural network based image classification using small training sample size
arXiv: VERY DEEP CONVOLUTIONAL NETWORKS FOR LARGE-SCALE IMAGE RECOGNITION