Visit

A Cornell research group has devised a way to generate an image of multiple people, incorporating the poses of all individuals to inform the image generation and make the scene more believable and accurate.

Read more on the Cornell Chronicle website