As mentioned in #17, the model code for SDC-Net in the sdcnet branch does not take use the spatially displace convolution block. Could you please elaborate on how to add the layer on the model sdc_net.py?
In the current code, I don't see how to use the autoencoder to retrieve the transformation parameters defined as input to the SDC layer network.
As mentioned in #17, the model code for SDC-Net in the sdcnet branch does not take use the spatially displace convolution block. Could you please elaborate on how to add the layer on the model sdc_net.py? In the current code, I don't see how to use the autoencoder to retrieve the transformation parameters defined as input to the SDC layer network.