Basically I have many X-ray images of ( human body and the liver within the body). Assume I can convert each of these image to a pair of images (the contour of body shape, and contour of liver shape, in the same coordinates system about their size and place). My purpose is to train some image generator or enhancer with these pairs of images, so that I can give it a new contour of a body shape, it can generate the contour of the corresponding liver in the right place of the body shape, so that I can add these 2 contour images together to “predict” where the liver would be given the contour of the body. This can be useful in medical treatment or even operation, particularly save some cost of medical treatment.
I have explored what GAN can do with quite a few youtube tutorials so far, and currently cannot think of a good way to use GAN for my purpose.
Thanks a lot in advance for the help!
Guiwen