Realistic Synthesis of Humans in Complex and Crowded Scenes

Nikita Kister (Ph.D. Student)

Recent advances in deep generative models have led to an unprecedented level of realism in synthesizing human faces and human avatars, reposing humans or swapping their identities and appearance. However, little attention has been paid on synthesizing full body humans in their realistic environments, surrounded by various objects, interacting with other people and performing different tasks. The goal of this project is to investigate how to augment real-world scenes with synthetic humans.

Primary Advisor:	Gerard Pons-Moll (University of Tübingen)
Industry Advisor:	Anna Khoreva (Bosch Center for AI)
PhD Duration:	17 May 2022 - 30 April 2025