A paper says
However, annotations used as inputs to C-GAN are typically based only on shape information, which can result in undesirable intensity distributions in the resulting artificially-created images.
What does "shape information" mean here?
I am aware of basic concept of GAN(generative adversarial networks), though I don't understand what "shape information" refers to.
I am aware the image above is an illustration of Image Segmentation. Would I think of the any one of the segmented area (red, green, blue) as a shape for a GAN?
Could someone please give a hint? Thanks in advance.