fenghetan9 / cmunext Goto Github PK

In the busi dataset,some images have multiple corresponding masks,for example ,benign (4).png have benign (4)_mask.png and benign (4)_mask_1.png.How do you deal with this problem?My Email:[email protected]

Hello my dear author, I am interested in your article and would like to get the code of the article. Hope to get in touch with you if it is convenient for you.My Email: [email protected]

I am interested for your works

I used MRI with nii file , how can I implement according to your project , please guide me .

thanks you

In the Ablation Study, what's the code of the "CMUNeXt Block" networks ( introduce the CMUNeXt block into the Reduced U-Net architecture ) in Table 3?
I change all fusion_conv to my DBLconv_block in the class CMUNeXt, for example, change self.Up_conv5 = fusion_conv(ch_in=dims[3] * 2, ch_out=dims[3]) to self.Up_conv5 = DBLconv_block(ch_in=dims[3] * 2, ch_out=dims[3]). The following is the DBLconv_block

class DBLconv_block(nn.Module):
    def __init__(self, ch_in, ch_out):
        super(DBLconv_block, self).__init__()
        self.conv = nn.Sequential(
            nn.Conv2d(ch_in, ch_out, kernel_size=3, stride=1, padding=1, bias=True),
            nn.BatchNorm2d(ch_out),
            nn.ReLU(inplace=True),
            nn.Conv2d(ch_out, ch_out, kernel_size=3, stride=1, padding=1, bias=True),
            nn.BatchNorm2d(ch_out),
            nn.ReLU(inplace=True)
        )

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

But its Params will be 3.01M , not 3.18 M as stated in Table 3.
So what's the code of the "CMUNeXt Block" networks in Table 3? My Email:[email protected]