I'm looking for a way to train an algorithm to reproduce an arbitrary transformation. Actually, I hope this is the right place to ask this question, it would be far easier to find a piece of software who already does the job, but I don't mind coding one (except for the fact that I have no experience with deep learning).
Think about deep fakes. The algorithm learns how to transform face A into face B and then applies this knowledge to any picture you feed it. I'm looking for exactly this, except more generic, maybe train it to learn how to transform a raven into a dove or something similar.
If there is no solution already coded for this, where should I begin?
