I have been using NiftyNet with various architectures, thank you all for your work. I am currently interested in using NiftyNet with GAN for am image synthesis problem (i.e. converting between modalities). It looks like the configuration file have to include the conditioning coordinates (as listed in the ultrasound demo example). Is it possible to use the GAN architecture without these conditions (just to have a generator and discriminator to convert between registered images).
Hi,
I have been using NiftyNet with various architectures, thank you all for your work. I am currently interested in using NiftyNet with GAN for am image synthesis problem (i.e. converting between modalities). It looks like the configuration file have to include the conditioning coordinates (as listed in the ultrasound demo example). Is it possible to use the GAN architecture without these conditions (just to have a generator and discriminator to convert between registered images).
Thanks a lot for your support