Using the neural-network Artbreeder, I’ve created caricatures of cartoonish people.
This is one of many experiments using Generative Adversarial Networks.
4-10 photographs of each person and and 2-5 cartoons were ‘embedded’ into the neural network for each result. Shown is the tree view of how the Sean Hannity caricature was made. Not show are the choices made in each step between ‘face’ (orientation / position / structure of the face) and ‘style’ (lighting, skin and hairstyle).
