Phenomenological models for the real-time animation of humans and animals

Contact: Marie-Paule Cani

Being able to animate visually realistic characters in real-time is a crucial challenge for the computer animation community. The applications go from the standard virtual humans used in crowd simulators or in video-games to the generation of the wild animals populating a natural scene, such as those we are developing in the EVASION group. We believe that generating real-time, easily controllable models for such characters cannot be done with the exclusive use of brute force, physically-based simulation. The aim of this work is to explore the combination of physically-based models with a phenomenological modelling of small, animated details such as cloth folds, skin wrinkles, dynamic hair and fur. The parameters of these models would be tailored by skeleton-based or physically-based animations performed at a much coarser scale. The models would have to capture the geometric constraints inherent to the elements we need to animate.

Prerequisite: C++, OpenGL, bases in Computer Graphics and Computer Animation

This PhD position is granted by the VISITOR European project.
Candidates should check their eliginility and apply according to the guidelines in: