Quoted By:
It's impressive how close it's getting with some of the recent ones. I've wondered about creating a neural network that can insert nipple pokies into images (anime or no). It seems difficult to gather analogous training data for a program like that, but it would be easy to find thousands of pics without nipple pokies, and thousands of pics with, and label them as such. The question is: is there a type of network that could make the leap from identifying a picture with/without one, to converting a picture without them to one with them?
