search

aurora95 / Keras-FCN

Keras-tensorflow implementation of Fully Convolutional Networks for Semantic Segmentation（Unfinished）
MIT License
650 stars 268 forks source link

I don't get the difference of the network structure between FCN_Resnet50_32s and AtrousFCN_Resnet50_16s #29

Closed wtliao closed 7 years ago

wtliao commented 7 years ago

Hi,

after analysing the code and comparing the model image, I don't find the difference of the network structure between FCN_Resnet50_32s and AtrousFCN_Resnet50_16s. The only difference is that AtrousFCN_Resnet50_16s used atrous_identity_block instead of the original in the last two identity blocks. But I don't get essential difference between atrous_identity_block and the original one either. As far as I know, the main difference between xxxx_32s and xxx_16s is a skip connection from the 16s downsampling layer to the 32s one. Am I right? Thanks.

ahundt commented 7 years ago

sounds about right. Atrous convolution changes the output dimension, it is a pretty straightforward difference.