| Rigid and chain-based models can cover coarse
full-body motion, but ignore subtle non-rigid deformations. Problem areas
are facial and lip-deformations, complex torso motion, muscle-deformations
on arms and legs, and shoe deformations. With this new project, we started
developing techniques that can pick-up those subtleties in 3D. We can recover
non-rigid basis shape models from monocular views in 3D (Non-rigid Structure
from Motion).
|
