search

xmengli / H-DenseUNet

TMI 2018. H-DenseUNet: Hybrid Densely Connected UNet for Liver and Tumor Segmentation from CT Volumes
526 stars 161 forks source link

Question about Upsampling layer input #34

Open Excelsiorwu opened 5 years ago

Excelsiorwu commented 5 years ago

Hello,I can not find 'long skip' in the Up-sampling process ,which your paper has mentioned. i wonder that the Up-sampling process only take the output of dense encoder without long skip from the front layers. in your paper image project just take X for input up0 = UpSampling2D(size=(2,2))(x) conv_up0 = Conv2D(768, (3, 3), padding="same", kernel_initializer="normal", name = "conv_up0")(up0)

xmengli commented 5 years ago

I attach two version of code. One is densenet and the other is denseunet, you can try two different version. The result difference is not very large.

Excelsiorwu commented 5 years ago

Thank you for your reply,I also find that 'long skip' do not play an important role in decoder process

xmengli commented 5 years ago

Which dataset do you use? I find it depends.