ProjectRVHu – Real Virtual Humans

Basic data

Acronym:
RVHu
Title:
Real Virtual Humans
Duration:
11/22/2018 to 4/22/2024
Abstract / short description:
Virtual humans are at the centre of diverse application areas such as medicine and psychology, virtual and augmented reality, and special effects in movies.
Furthermore, real humans can relate to human-like machines, which makes man-machine communication more natural. To achieve realism, all details matter: the facial expressions, the geometry of the body and its movements, the soft-tissue motion, the appearance and dynamics of clothing or the light reflecting on our body. All these are components that need to be modelled and perceived to the highest precision. An additional challenge is the uncanny valley, which hypothesizes that those digital humans which appear almost, but not exactly like real humans, elicit feelings of revulsion to the observers.

The proposed project RVHu aims at answering the following two inter-related research questions: How do we efficiently learn representations of humans without losing the detail that make us real? How can we train machines to reconstruct such representations from visual data?
Keywords:
digitization
Digitalisierung
machine learning
Maschinelles Lernen
3D reconstruction
3D-Rekonstruktion
computer vision
computer graphics
digital humans
3D deep learning

Involved staff

Managers

Department of Informatics
Faculty of Science

Local organizational units

Department of Informatics
Faculty of Science
University of Tübingen

Funders

Bonn, Nordrhein-Westfalen, Germany
Help

will be deleted permanently. This cannot be undone.